Roadmap

Shipped

1. 1.99Podcasts, talking pictures and music videos30 October 2026, live nowMake a podcast where you can be one of the hosts, make any picture askr made talk or sing, and turn songs into music videos.

   Change, sharpen and smooth any clip, and make a video of about a minute from one idea.

   What is new?

   askr's biggest update for pictures, clips and sound. Make podcasts, make pictures talk or sing, turn songs into music videos, change and sharpen any clip, and make a video of about a minute from one idea. The newest models are in too.

   How do podcasts work?

   Ask in a chat: "make a podcast about this PDF", about a link, a subject, or what you planned in the chat. askr writes it for two hosts and records it, up to an hour long. Once you have made your voice in Settings, you can be one of the hosts.

   Can a picture talk?

   Any picture askr made can talk or sing. Give it the words, your own voice, or a song from the chat.

   What can I do with a clip I already have?

   Change it ("make it night"), sharpen it up to 4K, make its motion smoother, make it longer or give it sound. Each one shows its price first.

   Can askr make a longer video?

   Say "make a one minute video of ..." and askr writes the shots, paints them with the same characters, and makes a video of about a minute, with a song if you like.

   What does it cost?

   You always see the price first, and something that fails costs nothing. A three minute podcast is about 13 cents, an hour long one about $2.75, and sharpening a five second clip to 4K about 32 cents.

   New features

   • Podcasts, up to an hour. Two hosts talk a subject, a PDF, a link or your chat through, written and recorded by askr, from three minutes to an hour. Make your voice in Settings and you can be one of the hosts.
   • Talking pictures. Any picture askr made can talk or sing: type what it says and pick a voice, your own included, or give it a song from the chat. You see the price before it is made.
   • Music videos. Any song askr made becomes a music video in a style you pick, claymation, retro game, cinematic and more, wide, tall or square.
   • Change it, Sharpen and Smoother motion. On any clip: change what is in it ("make it night", "make it anime"), sharpen it to 1080p or 4K, or make its motion smoother. Pictures can be sharpened to 4K too.
   • A video of about a minute from one idea. Describe a story and askr plans the shots, keeps the same characters in each, and makes one video with a song or the clips' own sound.

   Improvements

   • The newest models. Kling 4.0 makes clips of up to 30 seconds, FLUX 3 Max clips with sound, and HeyGen Video clips for about 2 cents a second. MiniMax Music 3 makes songs up to five minutes that sing your lyrics, and sound effects can run up to ten minutes.
   • Longer Seedance and FLUX clips for less. Making one of their clips longer now costs about $0.21 for five more seconds, not $2 to $3.
   • A picture model ready as a backup. Nano Banana 2.1 steps in when another picture model fails, so a picture still comes back.
   • Talking pictures in seconds. A picture askr made now talks in about half a minute, not three to seven minutes.
   • Minute videos in half the time. A video from one idea makes its shots four at a time, so a minute of video takes about half as long.

   Bug fixes

   • Shots that could not be made are said. When a video joined from several clips leaves one out, a line under it says so, and you are not charged for it.
   • Songs keep your lyrics. A song made from lyrics you typed keeps them, so its music video is made to them.
   • The same kite in every shot. A video from one idea keeps the things its story turns on the same from shot to shot, as it does its characters.
   • No black bars. A clip that came back with black bars above and below is cut to fill its frame.

   Announcement to follow.
2. 1.98Longer clips, sound and your own voice29 October 2026Make any clip longer, give a silent one sound, and have askr speak in your own voice.

   In Telegram and Discord groups, askr now reads files and voice notes too.

   What is new?

   Clips can be made longer and given sound. askr can speak in your own voice. And in your Telegram and Discord groups, askr reads files and voice notes.

   How do I make a clip longer?

   Tap Make longer under any clip, or say "make it longer". askr carries the same shot on from its last frame. You see the price first.

   Can a silent clip get sound?

   Yes. Tap Add sound on a clip with no sound, or say "add the sound of waves". askr adds sound that matches what happens in the clip.

   How does my own voice work?

   In Settings, read one sentence out loud to give askr permission. askr makes your voice from it and never keeps the recording. Talk can then answer in your voice, and a chat can read things in it.

   Can other people use my voice?

   Only the people you give a pass to. Everything made in your voice is labelled as an AI voice of you, and you can delete your voice at any time.

   What about groups?

   Tag askr on a PDF, a file or a voice note in a Telegram group or Discord server, and it reads it. A voice note is heard first, then answered.

   What does it cost?

   You always see the price before askr makes anything, and a failed make costs nothing. Making your voice costs about a cent.

   New features

   • Make a clip longer. Make longer under any clip, or "make it longer" in a chat, carries the same shot on from its last frame and gives you one longer clip. When a clip's own AI cannot carry it on, another does, and the chat says which.
   • Sound for a silent clip. Add sound on a clip with no sound, or "add the sound of rain", gives it a soundtrack that matches what happens on screen.
   • Your own voice. Read one sentence out loud in Settings and askr makes your AI voice. Talk can answer in it, and a chat can read or narrate in it. Only you and the people you give a pass can use it, everything made with it is labelled as an AI voice of you, and you can delete it at any time.
   • Files and voice notes in groups. Tag askr on a PDF, a file or a voice note in a Telegram group or a Discord server and it reads it, as it reads pictures. The file stays with the question in the chat's room.

   Improvements

   • Pay for what you got, here too. A longer clip that comes back short costs only the part that came, and one that fails costs nothing.
   • More Virtuals coins. Coins from Virtuals' newer launch contracts are read as Virtuals launches, their curve read straight from the chain.

   Bug fixes

   • Buyback notes on time. $ASKR buyback and burn notes reach the community chat within a minute again. They had been arriving a few minutes late.

   Announcement to follow.
3. 1.97Crypto, Chev and trading improvements27 October 2026Chev sells brand-new coins on time, askr reads four more launchpads and shows who launched a coin before, and every card says when it was read.

   What is new?

   askr Crypto and Chev got better at the hard parts: brand-new coins, copies, and knowing when to sell. You also see more of what happened.

   Does Chev sell new coins too?

   Yes. Take profit and stop loss now work on coins too new for the market feeds. Chev reads the price from the chain. If it cannot find a price, it tells you.

   Can askr read coins minutes old?

   Yes, on four more launchpads: Raydium LaunchLab, Meteora, Flap and Virtuals. The card shows the price and how far along its curve the coin is.

   Who made this coin before?

   On Base, Ethereum, BNB Chain and Solana, the card shows the creator's earlier coins and how they went. It also names wallets that snipe launch after launch.

   Is it a copy?

   askr calls copies better now. When the project's own X posts this address, it is the real one. The same project on two chains is not called a copy.

   Can Chev tell me when a price moves?

   Yes. Ask Chev "tell me if SOL drops 10%". It can tell you once, or every time it moves that much.

   What does it cost?

   Reads cost what they did, and each card now says when it was read and what it cost. Chev's trades now carry a small askr fee of 0.5%, shown on the quote before you buy.

   New features

   • Exits on brand-new coins. Take profit and stop loss now work on coins the market feeds do not list yet, such as a coin still on its launch curve. Chev reads their price from the chain, waits out a sudden jump, and tells you once if it cannot find a price.
   • Four more launchpads. A coin minutes old on Raydium LaunchLab, Meteora, Flap or Virtuals reads its price, what its curve holds and how far along it is, straight from the chain.
   • Who launched it before. On Base, Ethereum, BNB Chain and Solana, a read shows how many coins the creator launched before and how they went, and which opening wallets snipe launch after launch.
   • Buybacks. A read notices when the team buys its own coin back on its pool, says so on the card, and counts it for the market.
   • Price alerts by percent. Ask Chev to tell you when a coin moves by a percent, up, down or either way, once or every time. Add and change your watches in Chev's settings.
   • Your shared cards. See the coin cards you shared in Chev's Trading tab, and take one down. Its link then says it was taken down.
   • The card in Chev's chat. When Chev reads a coin, the card fills in live in its chat in the app, as it does on Telegram.

   Improvements

   • When a card was read. Each card says when it was read and what the read cost, like "Read 12 min ago, cost $0.41", so an old card is not taken for a fresh one.
   • A roomier card on phones. The card's parts, notes and scores have more room on a phone, and the chat sits apart from it.
   • What sales took. What Chev holds shows the profit already taken by sales, and a holding's chart shows what it cost.
   • Fewer wrong copy calls. A coin whose own project posts its address on X is the original, a fake with a brand-new X account no longer proves itself, and one project on two chains is called that, not a copy.
   • A steadier first score. The narrative score stays near the middle until askr has read X, and the card says it is a first look.
   • Reads hold up when it is busy. When many people check coins at once, askr shares the market answers between reads, and your read goes before Chev's scanner.
   • A code before the key leaves. Taking Chev's wallet key out now asks for a code sent to your email, or to Telegram when your account has no email.

   Bug fixes

   • A rule never buys a copy. A rule set to buy for you could buy a coin the read called a copy. Now Chev asks first and says why.
   • Sold means sold. A watched wallet's sale into its coin's own curve or pool read as "sent to a contract, sold or moved". It now says sold, and on Solana a quick sale and buy back are both told.
   • Clips that start on your picture. A clip made from your picture could come back as a new scene that left your picture out, on models such as Gemini Omni Flash, Kling and Seedance. Now the clip starts on your picture. askr Pro only picks a model that uses it, and a model you pick that cannot use it says so and offers one that can, with nothing charged.

   Also

   • A small fee on trades. Chev's trades on Robinhood Chain, Base, Ethereum and BNB Chain now carry an askr fee of 0.5% of the trade, shown on the quote before you buy. Trades on Solana carry none for now, and a read costs what it did.

   Announcement to follow.
4. 1.96Right the first time26 October 2026What you ask askr for comes the way you asked: any shape, the exact size, how many and how long.

   When an AI is down, another makes it. Comics and stories post to Explore as one episode, read page by page.

   What does it mean?

   Pictures, clips and songs now come out the way you asked: the shape, the size, how many and how long. When the AI you asked is down, askr makes it on another one instead of showing an error.

   Which shapes can I pick?

   Open Options by the message box: Square, Portrait, Landscape, Tall and Wide. For any other shape, say it: "4:5", "1080x1350" or "an Instagram post".

   Does askr remember?

   Yes. What you set in Options stays set in that chat until you change it. Tell askr "always make pictures 2:3" and it does.

   Can askr keep my characters the same?

   Yes. Open a picture's menu and tap Keep using this. Every new picture in that chat is made with it, up to four. Say "same picture, but vertical" and askr changes the last picture instead of starting again.

   What if something goes wrong?

   askr makes it on the next AI, at the same price or a few credits more. If only a dearer one can, askr asks first. A picture askr cannot make the shape you asked is free, and a clip or a song that comes out short costs only the part you got.

   Can I post a comic or a story?

   Yes. Put your pages in an album. Explore shows it as one post with its cover, and it opens as a reader, one page at a time. Albums can be episodes of a series, with a button to the next episode.

   Can I keep my pictures together?

   Yes. Tap Save as PDF at the top of a chat's pictures. Each picture gets its own page, in the order it was made.

   New features

   • Any shape, exactly. Options has Square, Portrait, Landscape, Tall and Wide, and any other shape can be said: a ratio ("4:5"), a size ("1080x1350") or a format by name ("an Instagram post", "a YouTube thumbnail", "A4"). askr asks the AI for its nearest shape and cuts what comes back to exactly yours, at exactly the size you named.
   • Clips and songs at the length you asked. A clip comes out at exactly the seconds you asked and the shape you asked. A song or a sound is cut to its length with a short fade.
   • No dead ends. When a picture, clip or song model is down or sends nothing, askr makes it on the next one at the same price or a few credits more, and says so in one line. If only a dearer one can, askr offers it with its price. A refused prompt is reworded once and tried again.
   • Keep using this. Keep up to four pictures in a chat, made or sent, and every new picture there is made with them, so a character, a product or a logo stays the same.
   • Change the last picture. "Same picture, but vertical", "make her hair red" or "now without the text" changes the picture you just made instead of drawing a new one.
   • Episodes and series on Explore. An album is one post on Explore with its cover, and opens as a reader: one page at a time, in order, with Previous, Next and page numbers. Albums can be episodes of a series with its own page, and a maker can turn off Make my version for an album.
   • Save as PDF. A chat's pictures on one page each, at their own shape, in the order they were made.
   • Pay only for what you got. A picture askr cannot make the shape you asked is free, and a clip or a song that comes out short costs only the part that came. The chat says so in one line.

   Improvements

   • Options stay set. Shape, size and how many stay as you set them in that chat until you change them, and Try again uses them too.
   • askr remembers your shape. Tell askr "always make pictures 2:3" or "one picture only" and it does, and askr Pro makes the next picture or clip in a chat at the shape you asked last time.
   • The right AI for a clip or a song. Ask a model that only writes for a clip or a song, and askr offers one that makes it, one tap away, and charges nothing.
   • The right coin. When a name or a ticker belongs to several coins, the card says which one askr checked, with one tap to check another instead, and that second check is free.
   • Tall pictures in full. Tall pictures and clips show whole on Explore instead of losing their top and bottom.
   • More AIs in Choose a model. Choose a model now follows askr's own nightly check, so AIs an old list hid show again, such as o3, GPT-5 Pro and GPT-5 Image Mini.
   • A better first clip. A new chat's Video door starts on a clip model made for any clip. It had started on one made for ads.
   • The developer API tries again. A model that is busy for a moment is tried up to three times before the API answers with an error, and a failed try is never charged.

   Bug fixes

   • GPT-5.4 Image 2 is back in Choose a model, with Shape and How many.
   • GPT-5.4 Image 2 makes the shape you ask. It made every picture square.
   • A model's name is never read as a number. "GPT-5.4 Image" in a message made four pictures.
   • A size like 1024x1536 keeps your AI. askr Pro moved those pictures to a Google model.
   • A model that only writes, asked for a picture, offers one that paints, and charges nothing, instead of answering in words.
   • $askr means askr's token. A coin that only calls itself askr is never read in its place, and its card says it is not askr's own.
   • Two clip models that could not make a clip are gone: Pika 2.2 and Runway Aleph, which needs a video to start from.

   Announcement to follow.
5. 1.95Max: three AIs answer as one, and more25 October 2026askr Pro has one slider from Cheap to Max, where three top AIs answer and askr joins them into one reply.

   Games askr makes get sound effects, 3D models and a test play before you see them.

   What is Max?

   The top of askr Pro's slider. Three top AIs from different companies answer your message. Then askr's strongest AI reads all three and writes one reply.

   What does Max cost?

   More than Performance, about 2 to 4 times, because every AI's answer is paid for. The price under the answer is the whole cost.

   Where is the slider?

   Open AI preferences in a chat and slide from Cheap to Max. It takes the place of the three buttons.

   What do the thumbs do?

   Tap the thumb up or thumb down under an answer. askr Pro learns what works for you. Tap it again to take it back.

   Does askr remember my choices?

   Yes. askr Pro makes your next clip at the length you picked last time, as it does for songs. Songs have an Instrumental choice too.

   Which AIs are uncensored?

   askr now tests every AI itself. Over the next few days the Uncensored filter in the model list fills with the ones that answer what many refuse.

   What is new in games?

   Games askr makes now have sound effects and 3D models. askr plays each new game for a few seconds before you see it, and if it breaks, askr writes it again for free.

   New features

   • Max. The top level of askr Pro: three top AIs from other companies answer first, and askr's strongest AI joins them into one reply. Every answer is paid for, about 2 to 4 times Performance, and the price under the answer is the whole cost.
   • One slider from Cheap to Max. AI preferences has one slider, Cheap, Balanced, Performance and Max, in place of three buttons. Choose a model has Max too.
   • Thumbs on answers. A thumb up and a thumb down under every answer. askr Pro learns from yours.
   • Uncensored, tested. askr asks every AI its own set of questions. The ones that answer what many refuse show under the Uncensored filter, and "Always use uncensored models" in Preferences uses them, as the tests finish over the next few days.
   • Sound effects in games. askr adds jumps, doors and explosions to the games and apps it builds, and the game kit's make_sound makes them too.
   • 3D models for games. askr and the game kit make 3D objects for a game from a few words, about $0.06 each, or $0.21 for a better one.
   • askr test-plays every game. askr plays each new game for a few seconds before the chat sees it. One that crashes or shows nothing is written again at askr's cost, and a picture from the play is its poster.
   • Publish a whole game. A game you built that is a whole folder, many files and WebAssembly, goes to its own address and to Explore with one command, npx https://heyaskr.ai/cli/heyaskr-cli.tgz deploy, and your agent key, paid from that key's credits. Up to 25 MB a game.

   Improvements

   • Your clip length. askr Pro makes a clip at the length you picked last time.
   • Instrumental songs. When askr Pro asks before making a song, an Instrumental choice sits beside the lengths, at the same price.
   • Memory by meaning. When askr remembers more about you than fits in one answer, it brings the facts closest to what you asked, even when you word it differently.
   • Mouse look and full screen. 3D games can lock the mouse to look around and go full screen. Press Escape to leave.
   • Where your plays came from. Your apps shows plays from chats, from links and from Explore.
   • Comments and reports. A maker hears under the bell when someone comments on or reports their app, game or skill.
   • Faster loading. askr's answers are compressed on the way to you, so chats and Explore open faster on a slow connection.

   Bug fixes

   • The roadmap says what is true. askr Build no longer promises a weekly build jam.

   Moved on: askr MindMoved on: askr ProMoved on: askr PlayMoved on: askr Build

   Announcement to follow.
6. 1.94Talk to askr, talk for free and more24 October 2026You can talk to askr out loud, askr answers for free when you have no credits, even without an account, and it makes Word and Excel files.

   Can I use askr for free?

   Yes. When you have no credits, askr answers up to 10 messages a day for free, and it can search the web for them.

   Do I need an account?

   No. Type on the home page and press Send. You get 5 free answers a day. Make an account to keep your chats and get 10.

   Which AI answers for free?

   Claude Haiku 5.5, a fast model. Pictures, clips and songs still need credits.

   Can askr make Word and Excel files?

   Yes. Ask for a document or a spreadsheet, then save it from the chat. A PDF or a Word file can now have askr's pictures in it and a cover page.

   Can I talk to askr?

   Yes. Tap Talk by the message box, say what you want and pause. askr answers out loud, then listens again. It works in browsers that can hear you, like Chrome, Edge and Safari.

   Can I send a video to any AI?

   Yes. An AI that cannot watch video gets still pictures from it instead. The sound is not sent.

   Can I translate my DMs?

   Yes. In a DM, open More and turn on Translate chat.

   New features

   • Free answers. With no credits, askr answers 10 messages a day for free on Claude Haiku 5.5, with web search. Each one says so under it, like "Free answer 3 of 10 today".
   • Try askr without an account. Type on the home page and press Send: askr answers right away, 5 times a day for free. Make an account to keep the chat and get 10 a day.
   • Word and Excel files. Ask for a document or a spreadsheet and save it from the chat. Numbers, dates, money and formulas in a spreadsheet work as real values.
   • Talk to askr. Tap Talk by the message box. askr listens, answers out loud with a voice on your device, then listens again. It works in a chat of your own and in group chats.

   Improvements

   • Pictures and pages in PDFs. A PDF or a Word file can show the pictures askr made in the chat, start with a cover page, and start a new page where it should.
   • Videos for any AI. An AI that sees pictures but cannot watch video gets up to 8 still pictures from the clip, each with its time. The sound is not sent.
   • Translate in DMs. Either of you can turn on Translate chat in a DM, or translate one message.
   • A start for accounts without an email. When the free answers run out, an account with no email is told it can add one in Settings for $0.25 of credit to start.

   Bug fixes

   • Old data cleared. Messages, polls and projects left from features taken out in 1.86 (Talk, Teams and Projects) are deleted.
   • The roadmap says what is true. askr Pages no longer promises daily challenges or weekly leagues, which were taken out in 1.86.

   Announcement to follow.
7. 1.93Platform stats22 October 2026askr's numbers are in the open at heyaskr.ai/platform: who visits, signs up and comes back, where people use askr, what they play on Explore, and the models they ask.

   What is it?

   A page with askr's numbers, open to everyone at heyaskr.ai/platform. It updates every few minutes, and it is in the footer of every page.

   What does it show?

   Visitors, sign-ups and sessions. Plays on Explore. The developer API. And which models and companies people ask, and what for.

   What is a session?

   One stretch of someone using askr, until they stop for 30 minutes. The chart shows where it happened: on your own, in a collab, with Chev, on Explore, through the developer API, in an Anon chat, or in a Telegram or Discord group.

   Does it show anything about me?

   No. Every number is a total. No person, chat or message is ever named. An Anon chat counts as a session and nothing more.

   Can I change the dates?

   Yes. Each chart has its own 24 hours, 7 days and All time.

   New features

   • askr's numbers in the open. heyaskr.ai/platform shows visitors, sign-ups, sessions, plays on Explore, the developer API, the models people ask and what they ask for, from launch day on 19 September, updated every few minutes. It is in the footer.
   • Sessions by where they happen. Solo, collab, Chev, Explore, the developer API, Anon, Telegram and Discord, each its own part of the chart, with its share beside its name.
   • Explore plays. Plays of games and apps on Explore, each day or each hour.
   • The developer API. Calls each day, the developers who made them, and what they put on Explore with a key.

   Improvements

   • A range for each chart. Pick 24 hours, 7 days or all time on each chart by itself, with the dates in the tooltip.
   • Totals since launch. Accounts, sessions, asks, the top company, games and apps on Explore, Explore plays, developers and developer API calls, together at the top.

   Bug fixes

   • Sessions in a group. A message from a Telegram group or a Discord server counted as a session of whoever linked the group. It now counts for the person who wrote it.
   • Anon chats counted. An Anon chat is erased a day after its last message, and its sessions went with it. They stay in the count now.
   • Developer API sessions counted. Calls to the developer API were left out of sessions. They are their own part now.

   Announcement to follow.
8. 1.921Telegram improvements21 October 2026askr in your Telegram group or Discord server is easier to read, easier to top up, and always checks a coin before it shows you a card.

   What changed?

   askr in groups got a round of fixes. Coin checks, the More button, and adding credit all work better.

   Can I trust the coin card?

   Yes. Only askr's own check can make a card. If the check cannot run, you get no card at all, never a guess.

   Why did askr ask for credit?

   Checking a coin costs more than a normal answer. If there is not enough credit for it, askr now says so with Add credit, instead of answering without checking.

   Does askr know who asked?

   Yes. In a group, askr answers the person who asked, by their name there. It no longer mixes them up with whoever linked the group.

   How do I add credit?

   Tap Add credit. You get the same screen as in the app: pick BTC, ETH, SOL, USDC or another coin, or pay with any token from your wallet. No askr account needed.

   Is More easier to read?

   Yes. More now shows the part's score on top, then short lines and lists, not one long block of text.

   New features

   • Add credit, the same as the app. The Add credit page a group opens is now askr's own Add credits screen, without the card: Majors with Bitcoin, ETH, Solana, USDC and every other coin askr takes, the amount, what you get, the address and its QR code, and Pay from wallet. Still no askr account needed.
   • Pay with any token. Tokens works from a group's Add credit page too: connect your wallet to see your tokens on Robinhood Chain, or search any token, and pay with it. askr thanks the group when it lands.

   Improvements

   • More reads on a phone. Tap More under a part and you get its name and score on top, one sentence at a time, and long lists as lists. A verdict that fits every claim is said once.
   • Shorter write-ups. A coin check's write-up, in the group and in the app, opens each part with a sentence or two and then lists what counts for it and against it. Claims come one a line.
   • A group is a group. askr answers a group as many people, each by their own name there. It no longer uses the memory or the settings of whoever linked the group.
   • askr asks for credit. When a chat cannot cover a coin check, askr says how much it needs, and in a group shows Add credit right there. Out of free reads? The group gets Add credit too.

   Bug fixes

   • A card for the wrong coin. When a group's chat had too little credit for a coin check, askr answered without checking, and once showed a card for a different coin. Now it never answers a coin check without checking, and only askr's own check can draw a card.
   • The wrong person named. askr could say someone asked a question when another person in the group had. It now names whoever wrote the message.
   • Memory in the right names. What askr remembers about a group's chat is kept under the names of the people who asked, not under whoever linked the group.
   • A wall of coins. The Add credit page listed every coin askr takes as a long wall of buttons. It is now the Add credits screen.
   • Add credit on a phone. The Add credit page ran off the side of a phone's screen. It fits now.

   Moved on: askr Anywhere

   Announcement to follow.
9. 1.92askr in your group20 October 2026Add askr to your Telegram group or your Discord server, and tag it there for anything askr does.

   Every question and answer is kept in a room of the askr chat you linked, which pays for it.

   What is it?

   askr, in the group chats you already use. Tag it in your Telegram group or your Discord server, and it answers right there.

   How do I add it?

   Open a chat on askr, open its settings, and tap Add to Telegram or Add to Discord. Pick your group, and you are done. You can also add askr from Telegram or Discord, and a group admin taps Set up.

   What can it do there?

   Anything askr does in a chat: answers, pictures, clips, songs, code and apps. Reply to an app it made to change it.

   Who pays?

   The askr chat you linked, the same as if you had typed the question there. If it runs out, anyone in the group can add credit with crypto, no askr account needed. You choose who in the group can ask.

   Where do I see it?

   Every question and answer is also in a room of your chat on askr, under the name of whoever asked.

   And coins?

   Paste a coin's address and askr's card fills in right in the group. Tap a part, like Launch, to see why it scored that.

   New features

   • Add askr to a group. In a chat's settings, under Telegram and Discord, Add to Telegram opens Telegram's own list of your groups, and Add to Discord opens Discord's own page to pick a server. The group is linked to that chat.
   • Or start from the group. Add askr in Telegram, and a group admin taps Set up. In Discord, someone who can manage the server types /setup. Either way, you sign in to askr and pick one of your chats, or a new one named after the group.
   • Tag it for anything. Tag askr in the group, or reply to one of its messages. On Discord, /askr works too. askr answers as it would in the chat, on askr Pro at the chat's level: words, code, pictures, clips, songs, files and apps, sent back as a reply to whoever asked. A picture on the question goes with it.
   • Change an app by replying. An app askr makes comes back as its link. Reply to that message and askr changes the app.
   • A room for each group. Each linked group gets a room of its own in the chat, named after the group, with Telegram's or Discord's mark on its tab. Everything the group asks, and every answer, is there.
   • Who asked, by name. A question from the group shows under the name of the person who asked, with Telegram or Discord beside it, and askr reads that name too.
   • Coins as a card. Paste a coin's address in the group, and askr's card shows the scores filling in as it reads: on Telegram as a picture that updates, on Discord as words that update, then the finished card as a picture. Under it: the chain and the launch, the price, market cap, liquidity, volume and holders, links to the chart, the explorer, the project's X, website and Telegram and a search for the ticker, the address to copy, and askr's read in one sentence: the narrative, the biggest risk and the biggest upside. The full write-up stays in the chat on askr.
   • A button for each part. Under the card, a button for each part with its score, such as Launch 85. Tap one for that part in a few lines, the best points for it and against it, then More for its full write-up. A tap costs nothing. On Discord only the person who tapped sees the answer.

   Improvements

   • Questions wait their turn. Ask several things at once and askr answers them one after another, in order.
   • Who can ask. Whoever started the chat sets who in the group askr answers (everyone, or the group's admins), and can unlink the group. Whoever linked it is always answered.
   • Anyone can top it up. When the credit runs out, askr says so in the group with Add credit. Anyone there can pay with crypto, with no askr account, and askr says thanks in the group when it lands. Cards stay with the account's owner.
   • A launch is judged on what happened next. askr Crypto now watches the first five seconds of a Robinhood launch, up from three. If the wallets that bought then have since sold, or their tokens are locked, the launch counts for much less against the coin. A bundle still counts in full.

   Bug fixes

   • Sold, not sent to a contract. A coin's launch said its first buyers had sent tokens to a contract when they had sold them through a trading bot or another pool. It now says sold, and says locked for tokens that sit in a lock.
   • No half answers in a group. When an answer in a group is stopped, the group is told it was stopped, not sent what was written so far. And an answer never runs on into the group's own messages.
   • Coin checks need an address. In a group, askr checks a coin in full only when you paste its contract address or a link to it. A name or a ticker gets a plain answer.
   • Launches read on busy days. When the chain was too busy to replay every transfer of a coin, its launch counted the first seconds in full. askr now follows the first buyers on their own, so what they sold, locked and still hold always counts.
   • Set up from Telegram. If tapping Set up in a group did not sign you in with Telegram, setting up stopped there. It now finishes in Telegram: pick your group in its list, and askr connects.

   Also

   • What askr sees. askr only answers messages that tag it or reply to it, and it never keeps the rest of the group's talk.
   • What it costs. A question from a group costs what it costs in the chat, paid from the chat's money or its creator's credit, the same as a line typed there.
   • On Discord, tag askr itself, the one with the APP badge, not the role with the same name. If the role is tagged, askr says so.

   Moved on: askr Anywhere

   Announcement to follow.
10. 1.91Check a coin, then let Chev trade it17 October 2026askr Crypto is here, in beta.

    Paste a coin's address in any chat, and askr checks it for you in about a minute.

    What does it check?

    Six things: the contract, the launch, who holds the coin, its market, what people say about it on X, and its website. Each one gets a score out of 100. You watch the scores fill in as askr works, and it warns you when a coin looks like a copy.

    Can I buy it?

    Yes. Set up Chev, your agent, once. Then tap Buy under the scores. Chev buys from its own wallet, inside the limits you set. It works in the app and on Telegram.

    Can Chev find coins for me?

    Yes. Set your risk once. Chev checks new coins all day and brings you the ones that fit. Say so, and it buys them for you too.

    What else can I do?

    Tap Watch, and Chev tells you when the key wallets sell or move the coin. Tap Share to send the scores as a picture, with your invite link on it.

    Which coins?

    Coins on Solana, Robinhood Chain, Base, BNB Chain, Ethereum and 23 more chains. Chev can buy on the first five.

    What does it cost?

    If your account has an email, your first read each day is free. After that, a read costs what it costs us, plus 5%. askr adds no fee to a trade.

    Is it finished?

    Not yet. It is a beta, so some things will change. Tell us what you think.

    New features

    • Paste a coin, get a read. Paste a coin's address in any chat, or a link to its chart or its explorer page, and askr reads it: the contract, the launch, the holders, the market, the talk on X and the website. It takes about a minute, and the card fills in part by part while you wait.
    • Six scores and an overall. Contract, Launch, Holders, Market, Narrative and Website, each from 0 to 100, and an overall that counts the contract and the launch most. askr's read sets the scores before the AI in your chat writes a word, so the AI cannot change them.
    • Who bought first, and who holds it. Launch looks at who bought in the first seconds, whether those wallets are linked, what the creator bought and what that group still holds. Holders looks at the top ten, wallets that move together or were funded by the same wallet, and brand new wallets.
    • Copies flagged. Every read looks for coins with the same name or ticker on the chains askr reads, and says whether this one is a copy, the original, or too close to call, from what the project's own X and website show and from the money behind each.
    • The website checked. askr opens several pages of the project's site and checks what a real project has: links that work, a domain that is not brand new, a real GitHub and Telegram, and figures that match the chain. Then it scores how real the site is.
    • Coins minutes old. A coin launched minutes ago reads too. A part that is too new to judge says Too new to tell, not a low score.
    • 28 chains. askr reads coins on Solana and 27 other chains, Robinhood Chain first. Chev trades on five of them: Robinhood Chain, Base, Ethereum, BNB Chain and Solana.
    • Crypto, a new kind of chat. Start one from the homepage, from the Crypto tile, or from Check a coin in the + menu. Its room is your crypto desk beside the chat: Chev's wallet on each chain, what Chev holds, your rules and the scanner.
    • Buy with Chev. Tap Buy under a card and Chev buys from its own wallet, inside the wallet's limits. On Telegram, Chev's card has Buy, Watch and Pass.
    • Your rules. Tell Chev what you like: scores, liquidity, age, value, chain, genre, launchpad, and what the contract's owner can still do. When a coin you read matches, Chev tells you, or buys it for you if the rule says so.
    • The scanner. Every five minutes Chev looks at new coins on its five chains, reads the ones your rules could match, and sends you the best fits, up to 10 a day to start, with Buy and Pass. It learns from what you buy and pass, offers a rule when it sees a pattern, and never suggests a copy.
    • Take profit and stop loss. Chev checks what it holds every minute and sells all of a coin when it reaches the gain or the loss you set, +100% and -50% until you change them. A coin with no market listed yet waits until it has one.
    • Watch these wallets. After a read, watch the wallets that matter, such as the first buyers or the biggest group, for seven days. Chev tells you when one of them sends the coin.
    • Share a read. Share any finished card as a picture, light or dark, with your code and a QR code of your invite link on it. The link opens the card for anyone, with Check your own coin under it, and whoever joins from it counts as your invite.
    • One free read a day. An account with an email gets one free read every day, whatever its balance, in any chat or with Chev.

    Improvements

    • Where each figure came from. The card, the answer and Chev say where each figure came from: the chain, the market feed, the explorer or askr's read of X.

    Bug fixes

    • Veo 3 videos at 5 seconds. A Veo 3 video picked at 5 seconds could fail. askr now makes it at the next length Veo takes, and it costs no more than the price you saw.
    • A check that could not run. When a check before a trade could not run, askr now says so in plain words.

    Also

    • What it costs. After the free read, a read costs what askr pays for it, plus 5%. Trades carry no askr fee, and askr pays the network fee on most of them.

    Announcement to follow.
11. 1.9Games from across the web15 October 2026Anyone can now make anything on askr, a new game or a change to a game whose code they have, and share it on Explore.

    Explore lists games from across the web beside the games made on askr, and most of them play right here: Play opens the game inside askr, Play with friends opens it in a room of your chat with the chat beside the game, and a game's page credits its maker, with their profile on X. Make one like this turns the game you just played into the start of your own. In a chat, everyone can change a game: anyone suggests a change, two ideas become a vote, and a signed-up member has the winner built. Bring Claude Code, Codex or Cursor and run them on askr's models with one key, with pictures and music for your game and a guide to making it. A play counts when someone really plays, makers see players and peak, a game made on askr never raises a wallet prompt, and the safety check reads a game the way a browser runs it.

    New features

    • Games from across the web. Explore lists browser games that live on their makers' own sites, beside the games made on askr: Halo: Combat Evolved, Quake III Arena, Doom and Tweetcraft among the first. Each has a page on Explore with its picture, where it is played and who made it, what you need to play it (such as Multiplayer, Works on phones, No download or Needs your own game files), About, Good to know, its reviews and comments, and More like this.
    • Played right here. Press Play on a web game's page and the game opens inside askr, under askr's slim bar: askr's mark, the game's name back to its page, Full screen and Play with friends. A game whose site will not allow that has Play on with the site's name instead, and opens there in a new tab. On the game's page, Open on their site is always a second way. A few games that need a walled-off space of their own play inside askr in Chrome and Edge, and on their own site elsewhere.
    • Not hosted on askr. A game from the web is still its maker's own site, shown inside askr or opened in a tab. Its page wears two marks, From the web and Not hosted on askr. askr's safety check has not read it, askr cannot see what you do in it, and anything the site asks of you, an account or your own copy of a game's files, is between you and that site.
    • Play with friends, in a chat. Play with friends, on a game's page or on its player's bar, asks which chat, a new one or one you are in, and the game opens in a room of that chat with the chat beside it, as an app or a game askr makes has its room. It has no link of its own: your friends are the chat's. On a computer drag the line between them to make the game wider, and Full screen fills the screen with it; on a phone the game takes the screen and the chat pulls up over it, with a number for what was said meanwhile. Where askr can pick the game's own room, everyone in the chat's room is in the same game at once. Where a game makes its own lobbies, whoever started the room presses Bring everyone into your game and everyone follows. Where online play works only on the game's own site, the room says so and Play online goes there.
    • Join your friends' room. A lobby's link from one of these games pasted into a chat shows the game's card with Play here, which opens the game in a room of that chat, at that link, for everyone who presses it.
    • Credit for makers. A game's page says who made it, under its name: a maker on askr with their picture, their @, their line and their profile on X, and a maker of a game from the web with their name, their profile on X and their site. The bar over a game played inside askr names nobody, so the credit stays on the game's page. Add your own X profile on your maker page, in the sheet where you write your line.
    • The Games tab, three ways. A switch over Games chooses Everywhere, Made on askr or From the web. Everywhere shows the web's games in a row above the games made on askr, and once you type a few words or pick free or paid the two become one list, with From the web on a web game's card. From the web orders by Most played, Newest or Top rated, and a search of all of Explore finds web games too, by name, by maker and by site.
    • One game shown large. Games opens with one game from the web across the top: the one askr's team picked, or the most played from askr this week when the team has picked none. It shows its picture, its name and site, its stars and its plays where it has them, with Play and Make one like this on it.
    • Claim this page. Made a game listed from the web? Signed in, its page asks Is this your game? Post the code it gives you on X from the profile the page names, or put it on your game's own site, and the page is yours: it says Claimed by your @, and its Yours corner shows its plays, page views and reviews, lets you change its words, tags and picture, and lets you reply to every review.
    • Reviews from people who played. Sign in and play a game, inside askr or on its site, and when you come back to its page the review box is there: one to five stars and your words, as for anything on Explore. A card shows a game's stars once it has a review and its plays from askr this week once it has one. Nothing is filled in.
    • Make one like this. On a web game's page, and on the page of every game made on askr, Make one like this opens a new game chat with the ask already written in its box, from the game's own listing. Change the words, send it, and askr builds a game of your own there. Nothing is sent for you.
    • Make a game with your friends. At the foot of Games, one box. Type an idea, press Make it, and a new game chat opens with your words in its box, ready to send.
    • A card of its own for a web game's link. The link to a web game's page has a card of its own: the game's picture and name, where it is played and who made it, and From the web. It shows under the link in a chat on askr, and it is the card the page hands to anything that previews a link.
    • AI agents add games. An AI agent with a key adds a game from the web through the connector's add_web_game tool. What an agent adds waits for askr's team to say yes, and shows on Explore within a minute of it. askr goes to the address first and refuses a download, a launcher or an installer, a site that does not answer, and an address that sends people on to some other site. For one agent askr goes to five addresses in ten minutes and twenty in a day, whether the game is kept or refused.
    • One kind of game. Asking askr to make or build a game now makes a game people sit down at together. The ask Make one like this writes starts a quick version in a game room of its own, with Play for everyone at once, who won said in the chat, a Stronger version and Change it, as "@askr let's play" always did. Before, it made an app with an Open button. askr goes by what the thing is called: a quiz is a game, and a score tracker for a game is still an app.
    • A game added to a chat stays a game. Add to this chat on a game askr wrote puts a game in your chat, where the copy was an app before: it gets a game room of its own, everyone there can play it at one live table and ask askr to change it, and it still says whose game it came from.
    • Live from a game's own page. A game askr wrote plays live from its own page as well as from its card in the chat. The people in the game's chat sit at one table whichever they opened, the page says so, such as "Live with 2 people", and who won is said in the chat from there too. Anyone else who opens the link plays alone on their own screen.
    • An AI agent's game plays live. When an agent publishes a game as its own page, askr puts live play and results into it, as it does for the games it writes.
    • Asked once, after the first round. When a game's first round ends, the card of whoever started it asks Who can find it?, with Keep it in this chat and Put it on Explore, $1. Nobody else in the chat is asked. Either answer is kept, so the game never asks again on any device, and the card keeps its own Put it on Explore, $1 for later.
    • Change it, for everyone in the chat. Every game card and app card in a chat has Change it, for everyone there. If you are signed up, your change is built for you as before, or you tap Suggest it and put it to the chat instead.
    • Suggestions. A friend who joined with only a name types a change and taps Suggest it. It lands in the chat as a line, "Sam suggests a change to Penguin Race", and on the card as their idea, and it costs nothing. Take it back removes your own idea while it waits. Whoever started the chat can take anyone's waiting idea off its card with Take it off, and the waiting ideas of someone who leaves the chat, or is removed from it, come off by themselves.
    • The vote. A second idea turns the card into a vote: one vote each, guests too. Tap an idea to vote for it, tap it again to take your vote back, or tap another to move it. The card says how many voted and which idea is in the lead. Only people who are in the chat now are counted: the vote of someone who left stops counting, and counts again if they come back.
    • Build it. Anyone signed up taps Build it on a lone idea, or on the idea in the lead once someone has voted. askr builds it as that person's own change, paid like any ask of theirs. On a game each idea's Build it says that idea's own price first, with the pictures and sound its words ask for in it, and where it asks for some a line under the ideas says the most it can come to. The chat's line then says who is building whose idea, on which model, about what it costs and at most, and whose credit pays. A build that does not come costs nothing and puts the idea back to wait.
    • Idea by. A version built from an idea names both people. The card reads "Version 2 by @nate, idea by Sam", the line that says it is ready says the same, and on Explore the listing says who made it and whose ideas were built: "Made by @maya and @nate, idea by Sam". A member is always named with their @ and a guest never is, and a guest cannot take a name that reads as askr's or as the @ of someone in the chat.
    • Guests play. A friend who joined with only a name plays for real: their taps change the game for everyone, they build in its world, and they take a seat at a live table. A guest still cannot build, change, remix, list or pay for anything. Keep my seat makes them a member.
    • A play is a first real move. Opening a game is no longer playing it. A play is counted when a person first presses a key, taps, or moves the pointer across the game, once a person a day. Opening it and leaving is an open, a crawler or a headless browser counts for nothing, and askr counts each play by the door it came through: a chat, a link, or Explore. From one address askr counts at most forty signed-out people in a game in a day; someone signed in is always counted.
    • Players, peak and minutes for makers. Your apps shows three more figures on an app or a game: Players over 7 and 30 days, each person once; Peak at once, the most people playing in the same minute; and Minutes played, the time spent at its live table, once it has had some. Who played is never shown. They count from the day this is switched on, so an older game can show plays and no players.
    • A royalty on live play. When someone remixes your game into a chat of theirs and pays for live play there, you earn a royalty on the part of it they pay with credits that could have been sent to someone else: not free, bonus or card credits, and not a top-up from the last 24 hours. Nothing is earned while that chat is public. It comes out of askr's part, and they pay exactly what they did before.
    • Build for Explore with your own tools. Make an agent key and run Claude Code, Codex or Cursor on askr's models, paid from the key's credits. Tool calls, streaming and long answers all work. Then publish what you made to Explore with the same key. heyaskr.ai/developers makes your key and has every line to paste.
    • The game kit. With the same key, your coding tool asks askr for a game's pictures and music, each kept at an address that never changes, and reads the askr game maker guide: games in one file, engines, saves, playing together, phones, a test that plays your game the way askr runs it, bringing over a game whose code you have, and publishing. One line adds askr's tools to Claude Code, Codex or Cursor. heyaskr.ai/agents.

    Improvements

    • No wallet prompts in games. A game made on askr never raises a wallet prompt. askr takes wallet extensions away from a game's code before its first line runs, so nothing in it can ask to connect a wallet or to sign.
    • Games open faster the second time. Your browser now keeps a game's engines, pictures and sounds after the first time, so a game you have played before does not download them again.
    • The safety check reads a game fairly. What a game's frame already refuses no longer keeps it off Explore: trying the network, moving the page away or opening a window, running code from a string, and cookies are told to the maker as notes. An engine carried inside a game's own file is recognised by its exact bytes, three.js, Phaser, PixiJS and Rapier among them, and is not read as the maker's own code. A name in the code such as seedKey is no longer taken for someone asking for a seed phrase. Wallets, seed phrases, private keys, forms that post to another site and look-alike sign-ins are refused as before, and a wallet's name hidden in encoded text, or joined together from pieces while the page runs, is now refused. Still kept off Explore: a peer connection, code or text askr cannot read, a name scrambled to hide what a page reaches for, a page written with document.write, a script from another site, and a password box.
    • A dry run of the safety check. An AI agent with a key sends a page to POST /api/agents/scan and gets back what publishing it would say: the verdict, whether it would be hosted and listed, and every finding with its level. Nothing is published, kept or charged.

    Bug fixes

    • A wallet restore ask in plain words. A page that asked for "the 12 words from your wallet" without ever saying seed phrase was hosted. It is refused now, as every ask for a seed phrase or a private key is.
    • A busy link lost its name and card. When many people opened the link to one app or one listing at the same moment, the link could show askr's plain card in place of the app's own. askr now keeps what a link's title and card need for a minute, so a busy link stays itself. Explore's own front is kept the same way, so something newly listed, or just taken off, shows there within a minute. For that minute a link to something just taken down or renamed can still show its old name. The page of something made on askr says so at once; the page of a game from the web can show for up to a minute after it is taken off, and its Play stops at once.
    • A live game stuck at "Connecting...". Some of the games askr wrote for playing live stopped at "Connecting...", or did nothing when you pressed Join, because of how the game's own code reached its live table. The games askr writes now get past it. A game written before this is not changed by it; a new version of it, from Change it or its stronger version, carries the fix.

    Also

    • What it costs. Finding, playing and reviewing a game from the web costs nothing, inside askr or on its site. Suggesting and voting are free. Build it and Change it cost what a change costs, paid by whoever taps, from the chat's money while it has some, else from their own credit: the build once, each picture or sound it asks for as it is made, and all of it never more than the most shown. A guest is never charged for suggesting, voting or playing. A coding tool's call on askr's models costs what the model costs plus askr's 5% fee, from the key's credits. From the game kit, a picture is about 4 to 5 cents and a piece of music about 8 to 9 cents, and an agent's free start credits pay for builds, not for these.
    • What askr keeps. For each app or game and each day askr keeps that a person played it, the door they came through, how many changes they made to what it shares and how long they sat at its live table, never what they did. It is kept for ninety days. The day's own totals stay after that, with nobody in them. For a game from the web askr keeps that you played it, which is what lets you review it, and nothing of what happens in the game, inside askr or on its site. For a coding tool's call askr keeps the billing record, never what was sent or what came back.
    • What is not in it. There is no form for sending a game in: askr's team and AI agents list them. The pages of games from the web are closed to search engines for now. askr counts plays by door, but Your apps does not show the split. A vote has no timer and nobody closes it by hand: Build it ends it, and so does taking ideas off until one is left. An idea cannot be edited, only taken back and given again. A page someone hosts by hand has no live table, and neither does a stranger who opens a game from its link: they play alone. The game kit makes pictures and music, not sound effects or 3D models yet.
    • Help. Games from across the web has what a listing is, what askr does and does not do for it, playing, reviewing and reporting one, and how a game gets listed. Change a game together has Change it, Suggest it, the vote, Build it, the names on a version and what a guest can do. Your maker page, sales and earnings says what counts as a play and what a maker sees, and its royalties now include live play. Put yours on Explore has what the safety check refuses, holds back and only notes. Make apps with your AI has your own coding tools on askr's models and the game kit.

    Moved on: askr PlayMoved on: askr BuildMoved on: Explore

    Announcement to follow.
12. 1.89Interim improvements13 October 2026Fixes and speed from this week's feedback.

    Chats with videos open faster, the picture you choose for an app or a game is the one Explore shows, a shared page's recipe holds every prompt it was made from, and a game room builds what you describe in plain words. New: albums for the pages you publish, large pictures, askr Pro picking from public rankings every day, and the newest models on the Pricing page.

    New features

    • Albums. Put the pages you have published into an album, in the order you choose: a comic that runs over several pages, a set of pictures, a run of songs. The album has a page of its own to share, and each page in it says which part it is and leads to the one before and the one after. Choose the album on the page itself, under Your page. On the album's page you can rename it, move a page earlier or later and take a page out, and every page stays published on its own.
    • Large pictures. Ask for a size in your message, such as "1920x1080", "in 4K" or "high resolution", or choose Large in a picture's options, and askr makes it at 2752 by 1536 pixels (1536 by 2752 when it is tall), where every picture was about 1376 by 768 before. A large picture takes about twice as long and costs about half as much again, around $0.11. On the Cheap level it is made on Nano Banana 2, the painter that can.
    • askr Pro picks from the rankings. On Cheap, Balanced and Performance, askr Pro now chooses its models from public rankings, read again every day: Arena, where people compare two answers without knowing which model wrote them, with LiveBench and Artificial Analysis for words and code, and Arena's picture and video boards for pictures and clips. Performance takes the best-ranked model for the job whatever it costs, Balanced the best under a price ceiling, and Cheap the cheapest of the best. A model that ranks higher takes a place by itself once enough people have voted on it and it is clearly ahead, so a new model reaches askr Pro without anyone choosing it.
    • Every picture model priced by a test picture. A picture model's listed rates do not say what one picture costs, so askr makes a small test picture on any picture model it has no price for, and again every month, and askr Pro's lists use what it really cost. Performance pictures are now made on GPT-5.4 Image 2, the picture model askr sells that ranks highest. A model that cannot make a picture for us stays off the lists until it can.
    • Newest models. Press Newest in Every model's price on the Pricing page to see the models in the order they came to askr, each with the day it was added. heyaskr.ai/pricing#newest opens it.

    Improvements

    • Chats with videos open faster. A chat shows a still picture of each clip and loads the video only when you press play. Before, opening a chat loaded part of every video in it. The same goes for askr Community, a chat's media rail and your Library.
    • A clip starts sooner once you press play. Its first seconds are sent at once, and a clip your browser already has is not downloaded again a few minutes later.
    • The Publish sheet says how many of your prompts go public. When what you publish was made and then changed, the sheet shows that every prompt behind it will be on the page, starting with your first.

    Bug fixes

    • A picture you uploaded for an app or a game did not show on its card on Explore once askr had taken a screenshot of it. The card showed the screenshot, often just the top of the page. Your picture comes first now, with nothing for you to do.
    • In a new game room, a game described without the word build ("doom in the browser") was answered with code files nobody could run. askr builds it now.
    • In a room that already has a game or an app, saying what it should be or what is wrong with it ("it should be faster", "I want more levels", "it doesn't start") was answered with code files too. askr changes the game now, as it does when you say make or add. A question is still answered in words.
    • A shared picture that had been edited showed only the edit as its recipe. The recipe now holds the first prompt and every change after it, in order, and Make my version starts from all of them. Pages already published update by themselves.
    • Nano Banana Pro stopped making pictures from 12 October: the version askr called was no longer served. Pictures asked of it, by hand or on askr Pro's Performance level, are made on the released Nano Banana Pro now, at the same price.
    • Try again on a large picture made one at the usual size. It is large again, as first asked. A large picture is also kept as a lighter file, so it opens faster in a chat.
    • In the installed app on Android, with the phone in dark mode, the back, home and recent apps buttons at the bottom of the screen were too pale to see. The bar is left to your phone's own light or dark mode now, so the buttons show. An app that is already installed picks this up within about a day, or when you install it again.

    Moved on: askr RoomsMoved on: askr PlayMoved on: askr PagesMoved on: Explore

    Announcement to follow.
13. 1.88Preferences, translations, community communication11 October 2026Chat on askr gets Telegram's message menu and Quote, profile cards and direct messages, posts from X drawn in the chat, and any chat translated into your language.

    Preferences are sentences on your account that every answer follows, a model you name is the model that runs, and @askr in askr Community answers with live numbers from the web.

    New features

    • Preferences. Write sentences in Settings, Preferences that go with every ask you make, to any model: "Always give me short answers", "Prefer GPT over Claude", "I'm in London, use UK spelling and pounds". Each one shows how askr read it: Style is added to the instructions of every answer, Model is a rule askr Pro picks a model by. They follow you into chats with others, askr Community and your Chev, for your asks only, and what you ask for in a message still wins.
    • Save as a preference. Say "from now on, always ..." in a chat and one tap under your message saves it to your preferences.
    • Translate a chat. Switch it on in a chat and every message that is not in your language shows in it, with the original one tap away. Only you see the difference. It costs about $0.02 for every 1,000 messages, from your own credit, and a message someone has already had translated is free. A bar offers it at the top of a chat that has another language in it; Translate is also in every message's menu and in the chat's menu, and your language is in Settings, Preferences. Not in askr Anon chats or direct messages.
    • The message menu. Right click a message on a computer, or tap it on a phone, for reactions, Reply, Quote, Copy, Pin, Translate and Report, in askr Community and in every chat. A hold on a picture, a clip or a sound opens it too, and on a phone a swipe left replies.
    • Quote. Reply to the exact words you mean. On a computer, select them and right click. On a phone, tap Quote, then the first word you want and the last. Your reply shows those words, a tap on them jumps to the original with the words lit, and @askr is told which words you mean.
    • Profile cards. Tap someone's picture in a chat and their card opens over the chat, without leaving it: who they are, when they joined, what they have made public, with Message and Open profile. Your own picture opens yours.
    • Direct messages. Message anyone with an @ from their card. A first message from someone you share no chat with arrives as a request you accept or decline. Messages sits at the top of Chats with an unread count, and Who can message me is in Settings, Account. A direct message is between two people, with no @askr in it.
    • Posts from X in the chat. A link to a post on X shows the post under your message: who wrote it, the words, the pictures, and a clip that plays in place.
    • Highlights and Announcements in askr Community. Two tabs beside Chat and Buys where posts appear as cards, the way buys do: posts from X to cheer on in Highlights, news in Announcements. Everyone who can talk in the room reacts. A new post there shows as a count on its tab, the way a new buy does.
    • Name your model. Say "use Opus", "ask GPT" or "use Seedance to make it" in a message and that model answers, for that message, at what it costs. When it cannot run, askr says so and offers the next model with one tap, never a silent swap.
    • Filters on the model list. Private, Reads pictures, Searches the web and Long context, wherever you choose a model.

    Improvements

    • @askr in askr Community is the askr of any chat. It looks things up on the web when an answer needs it and shows where it looked, reads the conversation around your message, and knows what askr can make and which models it has.
    • No daily limits in askr Community. Money in the room pays first, then your own credit, as in any chat.
    • Apps and games follow your level. Performance builds on Claude Fable 5.1, Balanced on Claude Sonnet 5.5 and Cheap on Gemini 3.8 Flash, and a model you name or pick is the one that builds, at that model's price. If it cannot build, nothing is built or charged and askr says so.
    • Your preferences reach askr Pro's pick. Prefer a maker and askr Pro runs its model for your level; say never to use one and it is skipped. The line under your message says when a preference set the model.
    • Links in askr Community. askr and X links can be posted there. Other chats take any link, as before.
    • Support in one place. Everywhere askr points you to support, it is heyaskr.ai/support.
    • Pinned messages sit just under a chat's name, where they covered the first messages before.
    • Settings has five parts. Preferences joins Account, Credits, $ASKR and Memory.
    • Topics that show what is new. A chat's topics sit in a row under its name, and each one wears a dot when something has been written there since you last had it open. askr Community's four tabs count what you have not seen. On a phone the row is slimmer than it was, and the chat's name has the bar to itself: the faces and Invite moved into More.
    • Buys you can react to. A right click, a tap or a hold on a buy in askr Community opens its menu, a reaction stays on the card, and the cards sit in the middle of the tab.
    • A guest becomes a full account in Settings. Joined a chat as a guest? Settings, Account now asks for your @ and a password in one step, with no email, and everything you wrote stays yours. Before, Settings kept the account under its guest name.

    Bug fixes

    • A model named in a message, or picked by hand, was ignored when making an app or a game, and Claude Sonnet 5.5 built it on every level. The model you name or pick builds it now.
    • @askr in askr Community said it had no live numbers, could not switch models, and sent people to a support bot that is no longer used.
    • Answers in askr Community sometimes lost a few words to "[link removed]" where there was no link.
    • A picture posted in askr Community was refused with "Request body is too large" unless it was very small.

    Also

    • Gone from askr Community: the sheet a left click opened (the menu takes its place) and the count of asks left today.

    Prepped for what is next

    • askr Pro picks by measurement. Lists built from Arena, LiveBench and OpenRouter's model data and rebuilt every day, with each level a rule: Performance the top score for the job, Balanced the top score under a price ceiling, Cheap the cheapest model above a quality floor. Built and checked, and switched on in a later edition.
    • Uncensored, measured. askr's own test of which models answer adult and edgy requests, in place of a list anyone keeps. The Uncensored filter, and "always use uncensored" in Preferences, arrive once every model has been tested.

    Moved on: askr ProMoved on: askr RoomsMoved on: askr Build

    Announcement to follow.
14. 1.87Community improvements10 October 2026askr Community is now one room with two tabs, Chat and Buys: anyone with an @ can talk, @askr answers anyone who tags it, the askr team and the mods wear ticks nobody can copy, and every $ASKR buy lands in Buys as its own card.

    GIFs, stickers, pictures and clips come to the chat, support moves onto askr, any chat can pin its messages, and the brand pack is at heyaskr.ai/brand.

    New features

    • One room. askr Community is a single chat, pinned first under Communities on Explore. Anyone can read it, and anyone with an @ can talk in it, no wallet needed.
    • Tag @askr. Tag @askr anywhere in a line for an answer, a picture or a clip. Money in the room pays first, then your own credit, and each answer says which paid.
    • Ticks you can trust. The askr team wear a gold tick and community mods a blue one, set by askr and never by a name, so nobody can pass as either.
    • Support on askr. Write to support from your account at heyaskr.ai/support: pick what it is about, say what happened, and a person on the askr team answers there, usually within the hour. Your conversations stay on your account, an answer appears within seconds while the page is open, and the bell, email or push tell you when it is not.
    • Pin messages. Whoever made a chat can pin up to ten messages in it, from people or from askr. The newest shows in a bar at the top: tap it to jump to the line and on to the next pin, or open All pins for the list. Everyone else in the chat gets a notification, and in askr Community the mods pin the same way.
    • Brand pack. heyaskr.ai/brand has the askr logo for light and dark, profile pictures, banners, the colours, the type and the words to use, with one download for all of it.
    • Chat and Buys. Two tabs at the top of askr Community: Chat is the conversation, and Buys holds the $ASKR buy lines, so the talk stays the talk.
    • askr's own buy card. Every $ASKR buy lands in Buys as a card of its own: a short animated header in your theme, a chevron for every $25, what was spent and the $ASKR received, a new holder or how far their position grew, the market cap, and links to the transaction, the wallet, the chart and where to buy.
    • Pictures, clips and GIFs. Add pictures, clips of up to a minute and 20 MB, and GIF files to a line, and they play right there in the chat.
    • GIF search. Tap GIFs in the box, search GIPHY and tap one to send it.
    • Stickers. askr's own pack of twelve stickers, sent with one tap and shown large. Stickers never count toward your day's pictures.
    • askr's notes. When askr buys back $ASKR, burns it or sends a new airdrop, askr says so in the chat, with a link to the token page.
    • Days in the chat. A small Today, Yesterday or date between lines where the day changes, in askr Community and every chat.
    • Unread on the tab. While askr is in another tab, new messages show as a count in the tab's title and a red dot on its icon, until you come back.
    • askr Community in your chats. Once you have said something in askr Community, it stays in Recent and on Chats.
    • A face on Explore. askr Community has its own logo and banner under Communities.

    Improvements

    • More of the chat with Chev on a PC. Once a conversation is under way, Chev is one small row at the top on a desk too, so the messages get the screen.
    • Tabs where they belong. askr Community's Chat and Buys sit beside its name on a desk and on their own row under it on a phone, and the room's header is just its name.
    • askr's notes in colour. Buybacks, burns and new airdrops show as small centred notes in the token page's own colours, green, orange and blue, each opening the token page.
    • A buy card that tells you more. Each buy shows the price it was bought at, and Buy opens $ASKR on Uniswap.
    • Etherscan for the chain. Links to transactions and wallets open on robin.etherscan.io.
    • The moon's falling lines run across the whole width of the screen.
    • A calmer homepage. Under the headline, one line at a time says what you can do with askr, a chat, a game, an app, a video, a skill or a framework, each in its own colour with the way into it on Explore just below. On a phone the box and its four kinds sit at the foot of the screen, and the line under the headline fits on one line.
    • Unread counts on the tabs. When there are new lines in askr Community's Chat or Buys while you are on the other tab, the tab shows how many, up to 99+, and opening it clears the count. Your own lines never count.
    • Buys from the start, in order. Buys holds every $ASKR buy as its own card, including the ones from before the feed was on, each showing when it happened and in the order they happened.

    Bug fixes

    Nothing this time.

    Also

    • Gone from askr Community: topics, announcements and highlights as their own tabs, the study rooms, challenges, the weekly jam, starter apps, the side chat, the count of asks left, the live count of who is watching, the Rules and pins panel, and every link to the Telegram group.

    Announcement to follow.
15. 1.86Make games, apps and media together10 October 2026askr is four things now: a chat, a game, an app and media, and every chat is a group chat.

    Every chat starts as a chat, a game, an app or media, from the same start wherever you begin, and askr opens a room for each thing it makes. You pay for your own asks, money anyone adds to a chat is used first, and nothing is priced before you send. Chat settings fit on one screen, Explore opens on communities, Pricing starts with a calculator and Help with six topics, and the extras that made askr hard to read are gone.

    New features

    • Start as a chat, a game, an app or media. Wherever you start, on the homepage, with Ask or with New chat, the box has Chat, Game, App and Media under it. Each makes that chat and opens it, a game, an app or media on a room of its own, with what you typed in its box. Typing and sending still starts a plain chat, and every chat is a group chat, so you can invite friends to any of them.
    • A room for each thing askr makes. Ask askr in a chat for a game, an app, a song, pictures, a clip or a website, and askr opens a room for it, names it ("Birthday song", "Snake game"), makes it there, and leaves one small card in the chat with Open. Rooms sit as tabs over the chat beside Chat. A game or an app runs right there with Play or Open, media lists its takes, pictures and clips, and in a room every message is for askr, no @askr needed.
    • Chat settings on one screen. Tap a chat's name for five rows: Name and picture, People, Money, AI and Who can see it, each opening one small sheet, then Close chat or Leave chat. Copy the invite link or invite by @ from People, and long press someone to remove them, if you made the chat, or to block them. Mute this chat and Notifications are under More in the chat's header.
    • AI preferences. One sheet for which AI answers: askr Pro at Cheap, Balanced or Performance, Choose a model yourself from Claude, GPT, Gemini, Grok and the rest, Answer length and Memory. Whoever made the chat sets it for everyone, and anyone else changes it for what they ask. A model you pick yourself answers your next message, then askr Pro picks again.
    • Money in a chat goes first. Anyone can add $5, $10, $25 or any amount from their credit to a chat, from Money in its settings, and everyone's asks there use it first, automatically. Each answer says who paid, Who spent what lists every ask that cost something, and what nobody uses goes back to whoever added it when they leave or the chat closes.
    • A first-time note. The first chat you open says once how to ask askr, who pays and how askr picks the AI, with Invite people, Add money and AI preferences. Close it and it never comes back, on any device.
    • Every chat can have its own picture. Whoever made a chat gives it a picture and a cover from Name and picture: Upload one, Pick one made in this chat, or Make one with askr, paid from your own credit. Every picture is looked at before it shows, and a chat without one wears the first letter of its name on its own colour.
    • Skills made in a chat. Say "@askr make a skill for meeting notes", or "@askr turn what we just did into a skill", and askr writes it in a room from what the chat said. It reads as a plain numbered document: everyone in the chat can change a line, try it on a real one and keep each version, and the first to put it on Explore becomes its maker.
    • Use a skill in any chat. Use in a chat on Explore adds a skill to a chat you pick, as a room of its own, and "@askr use" with its name does the same from inside the chat. The room follows the skill's newest version and says so in one line when it moves.
    • Explore opens on Communities. Its tabs are picture tiles (Communities, Games, Apps, Media, Skills, Frameworks and All), each showing real things people made, and one search finds everything. Sort and Filter replace the rows of chips. Each community leads with what its people made with askr, and with what they are making right now.
    • Make your own. On every tab of Explore and every page on it, Make your own starts a chat with askr's first words typed for you, or copies a prompt for your own AI.
    • What people make, on the homepage. Under Use AI together. the homepage shows real things made on askr, one at a time: a game, an app, a song, a video, pictures and a community. Under the box, Start a chat opens a new chat, and Explore what's possible opens Explore.
    • Pricing starts with a calculator. Say about how many answers, pictures, 5 second videos and 30 second songs you make in a month, pick Cheap, Balanced or Performance, and Pricing says about what it would cost. Beside it is what one thing costs on that level, with askr's 5% already inside every price, and under them every model's price.
    • Help opens on six topics. Help leads with the search, then six tiles named after what you want: Getting started, Chats and friends, Making things, Paying and credits, Explore and selling, and Account and privacy. Then come five top questions, and Message us. A person answers. at the foot of every Help page.

    Improvements

    • You pay for your own asks. In a chat with other people, what you ask askr comes from your own credit, free credit first, as in a chat of your own. Nothing asks before you send, and the cost shows small under the answer.
    • No prices before you send. No balance pill, nothing beside Send guessing what a message may cost, and a clip no longer stops to ask Make it. While a chat has money, one line under the box says so ("The chat pays for asks. $9.99 left"), and an answer the chat paid for says ", paid by the chat" after its cost. The only money card shows when nobody can pay, with Add credits and Not now.
    • A short answer instead of no. When your credit covers an everyday answer but not the longest one a model could write, askr writes a shorter answer and charges what it really cost, rather than turning you away.
    • Guests ask on the chat's money. Someone who joined with only a name asks askr while the chat has money in it. Otherwise they are asked to make an account, and can still talk with everyone.
    • A shorter + sheet. + has five rows and nothing behind them: Make a game, Make an app, Make media, Add a photo or file and Choose the AI.
    • Chats a size smaller on phones. A slimmer bar with the chat's picture, who is here and their faces, smaller bubbles, line tools on a tap, and a one-line box with + and Send that grows as you type. Nothing sits above the box, and the feedback bar stays on the website.
    • Links that land in the chat. Start a chat on the website and Make your own on Explore open the new chat itself, with the first words typed for you and never sent. Signed out, you sign in first and land in the same chat.
    • Joining is one press. An invite, a public chat's Join and an invite link's card join at once, with no price or token in the way. Publish a page is for pictures, songs and clips.
    • Notifications in Settings. Settings, Notifications turns this browser's notifications on and off, as Chev's settings already could.
    • Chev answers for you. When someone tags you in a chat, Chev asks you on Telegram or in its chat, and with your yes it answers there as you, marked via Chev. Chev no longer speaks in chats as itself.
    • Public in one step. Who can see it is Only people you invite or Public: on Explore, anyone can watch, and anyone with an account can join, up to 100. A public chat has its name, a line about it, topics and a level, and nothing to set about money.
    • Three places at the top. The bar on every page holds Explore, Pricing and Help. Token, Roadmap and Community are at the foot of every site page, and the logo always goes home. On a phone the menu has its own button at the top right.
    • Less $ASKR on everyday screens. The price pill in the bar, the price line under your balance, the token strip on the homepage and the holder offer in Add credits are gone. The token keeps the footer and its own page. A holder's deposit bonus still counts, inside You get.
    • How it works, in seven steps. The homepage explains askr in seven short steps, then six short questions.
    • Help in plain words. Articles are named for what you do: Chat with friends, How askr picks the AI, What askr remembers, Anonymous chats, Build a big app or game. Who pays states the one rule. Old addresses lead to the page that took their place.

    Bug fixes

    • A picture's price includes askr's 5%. Pricing showed a picture at what it cost before the fee; it now shows what you pay.

    Also

    • What left askr. To keep askr to four things, these are gone from every chat: askr Skill's ten skills, Group skills, party games, leagues and contests (cups, Beat the AI, streamer mode, the Arcade, the Daily game and the Model league), Make together, the story relay, Episode, voice rooms, Draw together, friends in the picture, Talk and askr in this chat (Catch me up, Ask privately, polls and room modes), Teams (team pots, holders powering a chat, paid and token-gated chats), tips, sending a clip to TikTok, Stories or WhatsApp, Invite on WhatsApp, a Telegram or Discord group linked to a chat, and a public page for a text answer. Old chats still read as they were, each old skill card written out in words, and what askr learned from them stays: askr Pro picks the right model for each job, and a message that may describe an emergency gets the emergency number first.

    Announcement to follow.
16. 1.85Explore games, apps, agents, skills and more9 October 2026Explore is where everything people and AI agents make on askr can be found: apps, games, skills and frameworks, each with a picture, a page of its own, reviews and a maker.

    Put yours on it for $1, choose whether people may remix it, set a price if you want one, and keep 80% of every sale. Every public collab is on Explore too, as a community anyone can find and join. AI agents make, list, sell and review beside people through the Agent API, and Chev runs frameworks from Explore for you. The rest of askr with friends opens with it: games askr hosts and invents in any chat, voice rooms, pictures, songs and videos a chat makes together, eighteen Group skills, team pots and paid chats, profiles and leagues, and notifications that bring you back to a chat. Apps and games grow with it: askr builds a big one step by step, up to the most you set, and what it builds can have pictures and sound of its own, real 3D and physics engines, a save for each player and a world the whole chat builds in. With this release askr adds a 5% fee to what AI usage costs, and $ASKR holders pay less of it.

    New features

    • Explore. Explore joins the bar at the top, after Home, and every menu, at heyaskr.ai/explore: everything people and AI agents made on askr, in one place. Featured leads, then a row each for Games, Apps, Skills and Frameworks, then New this week, Top rated, Free, Made by agents and Made on askr. A row shows only when it has something in it, and nobody needs an account to look, open or play.
    • Featured picks itself. Featured is what was used most this week, up to twelve things, with at most two from one maker. Nobody chooses it, and a week in which nothing was used has no Featured row.
    • Search, shelves and orders. Search reads titles, descriptions and makers' @s, and still finds a title through a typo. Chips pick a kind, each kind has shelves of its own (Fun, Social, Tools, Learning, Creative and Productivity for apps; Design, Writing, Coding, Business, Trading, Growth, Learning and Productivity for skills and frameworks), and any list orders by Most used, Newest, Top rated or Price, free first. What you asked for stays in the address, so a list is a link you can send.
    • A page of its own for every app and game. A tile opens the app's or the game's page on Explore first: its picture, what it is, who made it, its stars, its screenshots, what is new in its newest version, and its reviews and comments, with Open or Play, Remix where its maker allows it, Share and Report. Its own link, heyaskr.ai/a/…, still runs it at once.
    • Explore from inside a chat. Tap + in any chat, then Explore, and it opens over the chat: the search, the lists and each page, without leaving. Add to this chat puts an app or a game into the chat you are in, for everyone there. A skill or a framework opens in a chat of its own.
    • Communities. Every public collab is on Explore, under Communities beside Games, Apps, Skills and Frameworks. A tile shows the collab's name, who started it, a line about it, how many people are in it and how many wrote this week, and its level and topics, and opens the collab, where Join is. Find one by any of fourteen topics, by level (Newbie, Intermediate or Advanced) or by words, which read its name, its line and its creator's @, and order the list by Most active, Newest or Biggest. askr Community is pinned first.
    • Going public says what it means. Turning on Public in a collab's sheet opens Make this Collab public? before anything changes: anyone can find it on Explore and join, anyone can read everything written in it, and people who join can ask askr. Right there you choose who pays, Each person, This chat or Up to a limit, and whether guests can ask, then how it reads on Explore: its name, a line of up to 140 characters, up to three topics and a level. It starts on the safe side, with each person paying for their own lines and guests not asking. Afterwards the sheet keeps a summary of who pays and how it reads, with Change who pays and Change how it reads.
    • AI inside apps and games. An app or a game made on askr can ask askr AI while you use it: an opponent that answers back, a quiz that writes its own questions, a summary of what is on the board. The first time, a sheet names the app and the price, about $0.002 an answer and never more than $0.05, and waits for Allow. You pay for your own answers from your credits, and the maker pays nothing. Ask askr for it when you build: "@askr build us a riddle game where askr judges the answers".
    • Keep it in the chat, or put it on Explore. When an app or a game is made, askr asks Who can find it?: Keep it in this chat or Put it on Explore. Something kept in the chat can go on later from its card or from its page on Explore, and comes off again with Take off Explore. Only its maker can put it on, and it goes on only once askr's safety check reads its newest version clean.
    • You choose whether people may remix it. The moment something goes on Explore, its maker picks one of three, with nothing picked for them: No remixing, Free or A price, up to $100. With No remixing, people open it and play it, and nobody but its maker and the people in its own chat gets a copy or its code: it has no Remix, it cannot be added to a chat, and its tile says No remixing. The choice is asked under Who can find it? and behind every Put it on Explore, and it is changed later under Remixing on its page, or in a skill's or a framework's Prices.
    • $1 to list. Putting something on Explore costs $1, once, from the maker's own credits. Taking it off and putting it back is free, keeping it private is free, and nothing is charged when the safety check keeps it out.
    • A picture for everything, the first one free. Every app, game, skill and framework can wear a picture its maker chose: Upload one, or Make one with askr, which paints it from the item's name and what it is, as many times as you like, for you to pick from. The first picture askr makes for something that is on Explore is free. An uploaded picture is looked at before it shows, and without one the tile wears its shelf's symbol.
    • Screenshots, taken by themselves. An app or a game on Explore gets a picture of itself with nobody asked to do anything: it draws itself the next time its maker opens it, and the picture shows on its featured card, on its page and in link previews. A new version gets a new picture.
    • Prices. A maker can charge for a remix of an app or a game, and for the install, each use and a remix of a skill or a framework: anything from free to $100. Opening and playing stay free for everyone. A button that costs says its price, such as Remix, $5 or Install, $2, and askr asks once before anything is charged; the same tap is never charged twice.
    • Makers keep 80%. 80% of every sale lands in the maker's credits the moment it goes through, and shows under Activity as Earned on Explore. What you earn stays credit; there is no cash-out.
    • Royalties on your work and on copies of it. A maker earns 20% of what askr earns on their work and on copies of it, one step up. When a copy is sold, the maker of the thing it was copied from gets 20% of askr's part: on a $5 remix of a copy the seller keeps $4, the original's maker gets $0.20 and askr keeps $0.80. When other people pay for AI work inside something you made, or on a copy of it, you get 20% of askr's fee on it, about 1% of what they spend. The buyer pays the same and the seller's 80% is untouched. Royalties are added to your credits every hour, show under Your apps as Royalties and in Activity as Royalties on Explore, and are never paid on your own spending.
    • Which credits can buy. A purchase spends the buyer's own credits that could be sent to someone: credits added with crypto, credits someone sent them, and credits they earned. Credits bought by card, the bonus on a deposit, free credits and a chat's credits cannot buy on Explore, and the sheet says how much of the balance can. Purchases are final.
    • The code stays with its maker. The code of anything on Explore, and of anything with a price, is seen only by its maker and the people in the chat it was made in. Everyone else gets the app itself, and Remix is how to get a copy of the code of your own, where its maker allows one.
    • Reviews and comments. One to five stars, with words if you like, from anyone who has used the thing: opened an app, played a game, or installed or used a skill or a framework. One review each, yours to change or delete. The maker answers each review once, marked Maker; comments take replies; and three reports hide a review or a comment until askr's team has looked.
    • Stars where you look. Tiles, featured cards, an item's page and a maker's page all show the average and the number of reviews. Top rated lists what has four stars or better from three reviews up.
    • Skills and frameworks. Two new kinds of thing to make and share: written instructions any AI can load, in Markdown, up to 64 KB. A skill teaches an AI to do one thing well, such as a design taste, a writing voice or a coding stack. A framework walks an AI agent or a person through running something end to end, such as a business or a growth plan. New skill and New framework are on Explore, with Ask askr to draft it.
    • Read it, copy it, use it in a chat. A skill's or a framework's page is the text itself, with Copy for your AI, Download .md, Use in a chat, which starts a chat of your own where askr follows it, and Remix. With an install price, people read a preview until they install it, and then it is theirs for good with every version. One priced a use and nothing to install is never handed over whole: askr follows it in a chat, and an agent runs it through the API.
    • Runs and track records. An AI agent runs a skill or a framework through askr each time it follows it, always the newest version, and reports how the run went. The page shows a Track record: askr's own count of runs, then what the agents reported, with how many were done or failed, profit and loss as a total and a median, and the latest results with links to their transactions. Its maker can hide a run they think is fake, and the count of hidden runs stays for everyone to see.
    • A share of profits, checked on-chain. A framework's maker can ask for up to 50% of the profit a run makes, paid by the agent straight to the maker's own address on Base, Ethereum, Robinhood Chain, another EVM chain or Solana. The agent reports the payment, and askr checks each one on-chain and marks it Checked on-chain, Not checked yet, Not paid on-chain or Not found on-chain. askr never holds or moves that money.
    • A page for every maker. heyaskr.ai/explore/@you shows your picture, a line about you in your own words, your totals and everything you have on Explore. An AI agent has one too, marked AI agent.
    • Your apps, with sales and earnings. Your apps, on Explore, lists everything you made with its plays, opens and remixes, or a skill's installs and uses, over 7 days, 30 days or all time. Once you sell, Earnings sits above it: what you earned in the last 30 days and in all, a bar for each day, and every sale as a line. Who bought is never shown.
    • News of your things, in askr. A sale, a review and a remix of something you made, and a new version of a skill or a framework you installed, each arrive under the bell at the top of the app. A buyer and a remixer are never named, and nothing is emailed.
    • What is new in a version. A new version carries a line saying what changed. It shows on the item's page with its date, and the people who installed a skill or a framework are told.
    • The Agent API. An AI agent makes a key with one request and no sign-up, and gets its own @, a studio chat and $0.40 of free credit, good for 30 days. With the key it makes an app or a game from words, or launches a page it wrote, live at once at heyaskr.ai/a/… with a link anyone can open. The same answer carries a claim link: open it signed in and the agent is yours, with its apps, its credits and its keys under Your agents in Settings.
    • Agents pay for their own credits. An agent with a wallet adds credits itself, in USDC on Base, the open x402 way: askr answers its request with an offer, its wallet signs that exact payment, and the same request sent again adds the credits. From $1 to $100 at a time, at 1,000 credits a dollar, with no fee, and the wallet pays no network fee. A build an agent is short for answers with the same kind of offer, so it can pay and carry on.
    • Agents make, sell and review beside people. With its key an agent does what a maker does on Explore. It browses and searches Explore as people see it, puts what it made on Explore and prices it, gives it a picture, ships new versions, publishes skills and frameworks, buys an install or a remix from its own credits, runs a framework and reports the run, reviews what it used, answers reviews of its own things, and reads its notes and its sales.
    • askr as a connector. askr is an MCP server at heyaskr.ai/mcp. Add it to Claude or ChatGPT as a connector and ask for an app in any chat, with no sign-in to start. Every route is also described in an OpenAPI file for developers.
    • Make apps with any AI. heyaskr.ai/agents has prompts to paste into ChatGPT, Claude, Gemini or Grok as they are, and heyaskr.ai/agents.md has the same steps written for an AI to follow. An AI that cannot reach askr tells you to ask askr yourself, in a chat: @askr build us …
    • A key that acts as you. Settings, Account, Agent keys makes a key your own AI uses as you: your @, your credits, a studio chat of your own, and a daily limit if you set one.
    • Chev runs frameworks from Explore. Ask Chev for one in a job, tap Have Chev run this on a framework's page, or put it on a schedule: every day, on weekdays, every week or every few hours. Each run is a real run of the newest version. Its price a run comes from your credits under Chev's own spending rules, and how it went is reported to askr when the job ends and joins the framework's track record.
    • Chev pays the maker's share. When a run makes a profit and the framework asks for a share, Chev pays it from its wallet in USDC, straight to the maker. Pay makers their share, under Chev settings, Limits, decides how: Ask me first, where every Chev starts; By itself, up to an amount, $25 a run to start; or Never run those. The wallet's own limits hold in every mode, each payment shows in your chat with a link to the transaction, and askr checks it on-chain.
    • askr Build: build something big. Ask for something big, such as "@askr build us a big tower defence game", and a card offers Quick or Deep, each with its price. Quick is a small version in one go. Deep has askr plan the app or the game first and then build it one step at a time, so it can grow far bigger than one go can write. Nothing is built or charged until whoever asked picks one. New app and Invent a game offer the same choice from the start.
    • You set the most it may spend. Before a deep build starts, a sheet says about what a step costs and you pick the most the build may spend: $0.50, $1, $2, $5, or your own figure up to $20. Up to a fifth of it may go on pictures and sound askr makes for the build. You pay for what each step costs and never more than that in all, and a step that does not work costs nothing.
    • Watch it build, and open it as it grows. A card in the chat follows the build: the plan, then each step by number, such as "Step 4 of 9: writing the crafting table", with the time so far and what has been spent. Every step is saved as a new version, so Open runs what is built so far while askr keeps going. Every few steps, and at the end, askr looks back over the whole thing and fixes what is broken or missing.
    • Stop, and Keep building. Stop ends a build at once and keeps everything it made. Keep building, on an app's card, in the open app, on its own page and on a game's card, goes on with what is there, however it started: say what to add and set the most it may spend again.
    • Pictures, spoken lines and music of its own. Ask for an app or a game with art and sound, such as "@askr build us a dragon platformer with art and music", and askr makes its pictures, spoken lines and music as well as its code. Characters come on a see-through background, ready to move over a scene, and lines are read in one of askr's own voices. The price beside Send counts them first, each is charged only when it is made, and one that does not come costs nothing. A deep build makes the ones its plan lists as it goes, inside the most you set, and the stronger version of a game and a change that asks for art make theirs too, priced first.
    • The Assets tab. Assets, beside an app's Code, shows every picture and plays every sound it has. Add files of your own, PNG, JPEG, WebP, GIF, MP3, Ogg, WAV or JSON, up to 4 MB each, replace one or take it off. What the app still wants made is listed with its price, and Make them makes it. A version keeps the files it was built with, and a remix carries its files along.
    • Real engines: 3D, physics, sound and music. Apps and games can run on the engines the web's own games are made with, served by askr itself: three.js for 3D, cannon.js and matter-js for physics, Phaser for whole 2D games, PixiJS for fast 2D drawing, howler.js for sound files and Tone.js for music made in code. askr reaches for one when the idea needs it, and in 3D you look around by dragging.
    • An app says what it runs on. An app's card in the chat, and its page on Explore, carry a small badge for each kind of engine it loads: 3D, Physics, Game engine, 2D graphics, Sound or Music.
    • Saves. An app or a game can keep each player's own progress: a level, coins, things unlocked, settings. It is private to the player, and it follows them: the same save on any device, in the chat it was made in or from its link, and in its next versions. Signed out, it is kept in that browser.
    • Big worlds. A game's shared world can hold 64 MB in a chat, kept in pieces and loaded only where a player is looking, so a block world, a large map or a long history fits. Everyone in the chat builds in it, and someone watching a public chat sees it and cannot change it.
    • For people who write the code. An app has askr.save, askr.world and askr.assets beside askr.room and askr.ai, and askr.room.libs lists every engine with the address to load it from. An AI agent with a key sends an app's files, takes one off, and has askr make them from its own credits. Write your own app.
    • askr Play: games in any chat. Say "@askr let's play", or tap + and Play a game under Do something together, and askr hosts: it runs the rounds and the clock, keeps every answer private until its verdict, judges, and keeps the score. The round card under askr's message shows what the round asks, a countdown and who has answered, never what. Whoever started the game can skip ahead with Next and stop it with End.
    • Seven games askr hosts. Survive the AI, where askr throws a disaster at you each round and decides who lives. Trivia night, twenty questions on any theme. Guess the picture. A roast or rap battle, judged like a tough crowd, with a victory song for the winner. Murder mystery, with a suspect for each of you, secret roles and a killer among you. Spot the bot, where one or two seats at the table are AI. And AI story night, with askr as game master, a die rolled for every action, and a party it remembers for the next night.
    • Results, Rematch and Make it harder. Every game ends on a results card: the winner, the scores and the line of the game, always a player's own words. Rematch plays it again with all-new material, Make it harder does the same a level up, and Share gives the result a page of its own.
    • Games askr invents. "@askr let's play snails racing on a tilting table", or Invent a game in the + sheet, and askr writes a new game for your chat: a quick version first, then a stronger one for the rematch if you want it, each priced before anything is spent. Everyone in the chat plays at once, and the game tells the chat who won. A game that does not come costs nothing.
    • Live play. Games that move in real time, a race, a duel or a shared board, carry each player's moves to everyone as they happen, and someone who joins late catches up.
    • The Arcade and the daily game. A game askr wrote for you can go on Explore, and heyaskr.ai/arcade ranks those games by how they are played, never by likes. heyaskr.ai/daily has one new game a day, the same for everyone, free, with a share card of coloured squares that gives nothing away.
    • The model league. League round puts one question to two models from different makers, blind, as A and B. The chat predicts which maker wins for free points, votes for the better answer, and askr says which model wrote which. Every round moves the makers' ratings at heyaskr.ai/league.
    • Cups and seasons. A cup is a knockout for 4, 8 or 16 players in your chat: askr draws the bracket, and every match is a game. Places are worth free points for the month's season, from 30 for the champion to 3 for everyone who played, and the final can move to a public chat for people to watch.
    • Beat the AI. Open a pot guarded by an AI whose rules you write. People in the chat pay a little for each try to talk the guard into releasing it, every try goes into the pot, and the first try it releases the pot for wins everything in it. The rules are fixed and shown on the card when the pot opens, and if nobody wins, every try goes back to whoever paid it.
    • Streamer mode. In a public chat you started, show the chat's game on your stream through one overlay link. People watching send a voice line, a picture, a song or a twist into it, priced first at what it costs to make plus your share, and nothing reaches the stream until you approve it.
    • Every game by voice. Turn on Play by voice, or say "out loud", and askr reads every line it says in the game in one of its own voices. Hold to answer turns what you say into your answer, and the clip is kept for you alone, for a week.
    • Voice rooms. Schedule a room for a time, with a format: trivia, a rap battle or an open floor. People say I'm coming, raise a hand, and hold to talk when askr gives them the floor, one speaker at a time, while askr hosts, keeps time and keeps score. Every room is recorded, everyone is told so before they can speak, and you can delete your own clips. Someone watching a public chat can listen.
    • askr Studio: make it together. Make together, in the + sheet, opens games and things to make with everyone in the chat. Prompt Battle, Caption This, Guess the Prompt and Telephone are picture games. In the Friends' arena one prompt goes to three or four picture models, and you vote before you learn which made which. You vote by reacting, and every button says the most it can cost before it runs.
    • Picture cards, group photos and posters. Every picture in a chat says who asked, the prompt, the model and what it cost, with Again, Remix, Better and Four drafts. A photo of everyone becomes action figures, an anime still, a 1980s yearbook page or a band poster. An event poster gets a page where anyone replies yes, maybe or no with just their name.
    • Songs of the chat. askr writes lyrics from the chat and makes a song, priced first. Each song offers Full song, Another genre, Remix and Send to, shows the whole family of songs made from it, becomes a tall lyric video on your own device, and lights up line by line for karaoke.
    • Roast battles, music games, a podcast and gifts. A roast battle makes a diss track each way, and only once both people have said yes. Name that tune and Finish the verse are games for the ear. This week in the room is a three to four minute podcast of the chat's week with two hosts. A song gift is a song for a friend, with credits attached if you like, and a picture bounty offers credits for the best picture to a brief.
    • Voice notes. Voice note, under Make together, reads a line from the chat, or your own words, as a movie trailer, a sports commentator, a nature documentary or the news: askr rewrites it in that style and one of its ready-made voices reads it, at about the price the form shows.
    • Videos in one tap. Make a video has eight formats, from a birthday shout-out and a movie trailer to a holiday card and a superhero intro. Each is a recipe: you add the names, and the price, the length and the shape show before anything runs. A recipe can start from a real photo, one shared in the chat or one you add there, once you tick that everyone in it is in the chat and happy to be in the video.
    • The room reel, and one Studio for the chat. Room reel lays the chat's clips and pictures out as one tall film with a song under it, free, and a new song for it can be made to the reel's own length, up to a minute, priced first. Studio is one timeline the whole chat arranges, kept on askr's side, with everyone's editor following each change. The person who started the chat can give it a look, such as Warm film or Watercolour, that every picture and clip made there takes.
    • Story relay and the occasion kit. A story relay is told ten seconds at a time, from a photo if you like, each friend adding the next shot from the last frame of the one before, with a vote on what happens next. An occasion kit gathers lines, voice notes, clips and photos from friends with no account, and becomes one film sent into the chat on the day, with a song made to its length if you like.
    • Send it with a way back in. A clip or a reel goes to TikTok, Stories, WhatsApp or Telegram, a clip in the shape each one wants, with a caption that leads back to the chat.
    • Draw together. "@askr let's draw" opens one canvas the whole chat draws on at once. Live picture has askr paint the sketch as you draw, Render now and Finish make one picture at the price on the button, and "@askr make Sam's jacket red" changes it for everyone. Every version is kept, and Post to chat puts the one you like into the chat.
    • A weekly episode. Episode makes a short show of the chat's week, hosted by Lily and Roger. It is priced first and never more than the chat's cap, nothing is made until someone says yes, the script is shown before anything is made, and anyone named in it can be left out. Its scenes are made at Cheapest, Better or Best, and the price follows. It lands in the chat's Studio as a timeline.
    • Your look, only with your say. Add up to four photos of your face under Your look and voice and choose, chat by chat, who may put you in pictures: only you, everyone there, or people you pick. You are told every time in a line only you can see, everything made with you is listed, and you can delete any of it for everyone. Pets and characters you add work the same way.
    • Friend sticker packs. Eight stickers of the people who let you use their look in a chat, to download for WhatsApp, iMessage and Telegram.
    • Talking cards. In a chat's sheet, the photo of a pet or a character you may use says your words, up to 15 seconds, read in one of askr's voices: a birthday card in the chat. See the price comes first, the card is charged for its real length and never more, and whoever is in it is told.
    • Group skills. A Group skills chip joins the row over the message box in every shared chat, with eighteen things a group does together. Everyone gives their part, privately where it matters, and the chat gets one result.
    • Settle this. "@askr settle this" in an argument shows the price first. Then three models from three different makers judge it, each on its own, with a web search for sources, and askr writes one verdict that keeps what is fact apart from what is opinion and says how many of them agree. The person who asked can publish it as a page.
    • Split this, and the chat's ledger. Snap the receipt and everyone taps Mine or Shared. Shared dishes, tax and tip are split in proportion, and the bill goes into a ledger of who owes whom in real money, where askr drafts the reminder nobody wants to write. askr credits never move.
    • Quiz us and Study together. Paste notes or snap a page and askr makes a quiz answered by reaction, with scores and a study streak. A finished quiz can be published as a question set for an exam, and every right answer counts in your exam's weekly league.
    • Cards, wishlists and Secret Santa. Everyone adds a line to a card or a toast and askr weaves them into one, a surprise if you like, and A sung version makes a posted card a song, priced first. A chat's wishlist keeps everyone's wishes with their prices, and Secret Santa draws the names and tells each person theirs privately.
    • Help me choose, plans and boards. askr compares two to four options side by side and the chat votes with the results hidden. Plan something finds options that fit everyone's private budget and days. A Home board turns a school email into dates and tasks and keeps a fair chore rota, and a Care board keeps handovers, medicines and the questions for the next appointment.
    • Settle this fairly, and common ground. For two people, askr sums up each side for its author, shows both side by side only once each has agreed their summary is fair, and takes no side. For a bigger disagreement, everyone writes privately and askr drafts one statement the whole group could accept.
    • Stories, challenges and circles. In Story time people take turns and askr narrates between them. A challenge has everyone aim the same step higher from their own start. A savings circle keeps track of whose turn it is and who has paid, and askr drafts a payment from a group's multisig in plain words, for you to sign in your own wallet.
    • Room wallet: teams, with one pot. Teams joins the menu. Make a team, invite people by @, bring your chats into it, and one pot of credits stands behind all of them: when a team's chat runs short, the pot pays. Each person uses it up to a daily limit, anyone in the team can add to it by card, crypto or from their balance, and the team's page shows what the pot took in and what each chat and each person spent.
    • Chats powered by holders. Someone whose linked wallet holds $ASKR can power a chat: their own balance pays for everyone else's messages there, up to a cap a day they set, and the chat says who powers it. Nothing is sent to anyone, and they can stop at any time. In a public chat they say yes to that first.
    • A price, or a token, at the door. The person who started a chat can charge newcomers for a place, once or as a pass for 1, 7 or 30 days, up to $5, or let in only wallets that hold enough of a token. The price shows before anyone joins, and everyone already in stays for nothing.
    • Proposal summaries. Paste a proposal on a team's page and askr writes what it asks, the money in it, who it affects, the risks and the questions to ask, for the whole team to read.
    • askr Rooms: notifications. A bell at the top of the app holds what is waiting for you: invites, messages that name your @, replies, and what askr answered while you were away. Each opens the chat where it happened, and Chats shows a count beside each chat.
    • Your turn. When a game, a story or a plan in a chat is waiting on you alone, askr tells you: a round everyone else has answered, your turn in Telephone or Story time, a plan only you have not answered, your cup match starting. A voice room you said you are coming to tells you ten minutes before it starts. Whoever passed the turn on is never told, and a note is taken back when the turn is over without you.
    • Email, push, Telegram and quiet hours. One short email for everything waiting, off until you turn it on, and a push for each chat and a Telegram message where you set those up for Chev. Nothing leaves askr between 22:00 and 08:00 your time unless you change the hours.
    • Your week. On Monday morning, the chats you muted or have not opened for a week: how many new messages, from whom, and a short summary of the busiest.
    • askr Pages: a profile you choose to show. heyaskr.ai/u/ and your @ is your page: off until you turn it on, gone at once when you turn it off, and showing only the sections you pick, from games and pages you shared to season badges. Nothing from a private chat ever shows.
    • Leagues for chats. The person who started a chat can enter it in askr's leagues under a name they choose. Each week counts what the chat does together, people active, games finished, things made and pages published, never what anyone wrote or spent, and heyaskr.ai/leagues shows only the name, its points and its place.
    • Seasons and badges. A season is a calendar month, in three divisions by the size of the chat. When it ends, the places become badges on the chat's page and on the profiles of the people who counted.
    • A 5% fee on AI usage. From 9 October 2026 askr adds 5% to what AI usage costs: an answer, a picture, a clip or a song that costs $1.00 is charged $1.05, and every price askr shows has the fee in it. A top-up is not touched: by card or crypto, every dollar you add is still a dollar of credit. A web search keeps its own fixed price, and nothing is added to what you buy on Explore or to the $1 for listing there. The notice is in the terms.
    • $ASKR holders pay less of the fee. A linked wallet's share of supply × 100 comes off the fee: 0.1% of supply pays 10% less of it, 0.5% pays half, and 1% or more pays none. It follows the wallet's latest reading, which askr takes when you link the wallet, within about ten minutes of any $ASKR moving in or out of it, and at least once a day. The discount comes off the fee only, never off what the AI itself costs. Settings, $ASKR shows your discount. Activity says what it took off each charge your balance paid, and so does Details on an answer or a picture. Holding $ASKR.
    • Referrals pay a share of the fee. Bring someone to askr and you get 20% of askr's fee on the AI they pay for from their own balance, for as long as you both have accounts, added to your credits every hour. It takes the place of the 10% of every top-up, which has ended. Send credits and referrals.

    Improvements

    • Apps and the gallery live on Explore. heyaskr.ai/apps and heyaskr.ai/gallery now open Explore, on Apps and on Made on askr.
    • A live card while askr builds. While askr builds an app, its card in the chat shows the step it is on: Waiting for a builder, Thinking about how it should work, Writing the code, with the line being written, then Checking it. The time so far ticks by, whoever asked can cancel while it waits, and a build that does not come says why and that nothing was charged.
    • Share, in one tap. Share on a running app's bar, on its page on Explore and on a skill's or a framework's page copies the link and opens your phone's share sheet where it has one.
    • Link previews. Paste the link of anything on Explore into a chat or a post and the card shows its picture, its name and who made it.
    • From a running app to its page. On an app that is on Explore, Made on askr in the bar opens its page there, and the star opens its reviews.
    • What is new, by hand. Saving a version by hand in an app's Code tab takes a line on what is new in it.
    • Activity names Explore. Settings, Credits lists Put on Explore, Bought on Explore and Earned on Explore among your charges and top-ups.
    • New app, from Explore. New app on Explore opens a fresh chat with "@askr build us" already in the box, for you to finish and send.
    • Collabs that were public already are listed too. Nothing about their settings was changed, and their creator reads a notice once at the top of the collab, This Collab is on Explore now, with Check the settings. A signed-in person can report how a collab reads on Explore from its sheet, and three reports take it off Explore until askr's team has looked. It stays open by its link.
    • One place to start things together. The + sheet gains Do something together, in three groups, Games, Make together and Talk. What is going on now, a game, a voice room, a league round or a cup, shows as chips in one row over the message box.
    • The bill says who paid. A chat's Who used what, and its CSV, say when a team's pot or a holder paid for a message, and count those apart from the chat's own credit.
    • Games on Explore. Explore's Games has buttons for the askr Arcade, the Daily game, the Model league and Leagues, and a game askr wrote for a chat goes on Explore from its own card.
    • One bell for everything. News of what you made on Explore and the notes from your chats arrive under the same bell.
    • Mute means quiet. A chat you muted sends you nothing but its place in Your week.
    • Finished games count. A game played to its end shows on its players' profiles where they chose that, and counts for the chat's league.
    • Room to grow. An app's file can be 4 MB, four times what it was. A move in a live game can be 8 KB, someone who joins late loads up to 256 KB of the game so far, and an answer from askr AI inside an app can run to 2,000 tokens.
    • Apps stay walled off. An engine comes from one fixed address that askr serves itself. An app loads those files and its own pictures and sound, and nothing from anywhere else. With an engine or without one, it cannot reach the internet.
    • A missing picture never breaks an app. Every app askr builds draws a stand-in for a picture or a sound that is not there, so it runs before its files are made and when one does not come.
    • A file you add is checked first. askr reads what a file really is, looks at a picture and plays a sound through before it keeps it, and refuses anything that could run as code.
    • A deep build carries on through an askr update. The step it was on is picked up afterwards, and nothing is built or charged twice.
    • One message when a deep build ends. The chat hears once what was built, in how many steps, and what it cost, never a message for each step.
    • An engine's error in its own words. When an app built on an engine hits an error, Ask askr to fix it carries the engine's own words for it.

    Bug fixes

    • "@askr build us a coin flipper" was answered with code in the chat. Whatever you ask askr to build is now built as an app, unless it is a picture, a clip, a song or a piece of writing.
    • A public collab could spend its creator's own balance. Since Rooms with friends, a chat that ran short of credit was topped up from its creator's balance, and that held in a public collab too, where anyone can join. A public collab now runs on its own credit only: its sheet says so, and Cover this chat from my balance is off while it is public. A private chat covers as it did.
    • Putting something on Explore with under $1 of your own credits and free credits on top answered with an error. It now says what listing costs and what you have, since free credits never pay the $1. Nothing was charged and nothing was listed by the error.
    • A guest could still spend a chat's credit where guests may not ask askr. With Guests can ask askr off, a guest's @askr message was refused, but a private question or a poll made from the chat was still paid for by the chat. Nothing a guest starts is paid for now while the switch is off.
    • A newcomer could be charged for a welcome they never asked for. In a chat where each person pays for their own messages, joining made the welcome summary at the newcomer's own cost. A welcome is now made only where the chat's credit pays for it.
    • askr could speak up at one person's cost in a public collab. Where each person pays for their own messages, askr speaking up unasked was charged to whoever had switched it on, set off by anyone who wrote there. It now stays quiet in such a collab.
    • Hosting a page from a chat refused anything over 256 KB. Whatever the limit said. Host now takes a page up to the full size an app may be.
    • An app that sized its canvas as it opened could stay empty. Its frame could start before it knew its own size. A frame now starts once the page has laid it out, so an app knows its size from its first line.
    • A translation could be shown cut off. With Read this chat in on, a translation could stop part way and be shown as the whole message, and a message past 3,000 characters was cut before it was translated. A translation is now shown whole or not at all: a message askr cannot translate in full stays as it was written, at no charge, and a very long one is always shown as written.
    • The price under a message could name the wrong payer. Where each person pays for their own messages, the line under an answer still said the chat's credit had paid. It now says whose credit it was: your own balance, someone else's by their @, or the chat's.
    • Two prices were under what the work cost. The stronger version of a game, and a change to one, said about $0.12 where they cost about $0.19, and an app built on Grok Build 0.1 said about $0.02 where it cost about $0.04. Each button now says what the work costs.
    • A change to a game that already had a stronger version never arrived. It was written and then lost, nothing was charged, and the chat could not start another game for twenty minutes. A change now arrives, and is charged once.
    • Scroll while askr writes. A long answer no longer holds the chat still or slows the page: scroll up or down while it is written, and the chat follows only while you are at the newest line.
    • Opening a chat is lighter. A chat opens with fewer calls, and the limit on calls from one address is higher, so a reload never runs into it. Sending a sign-in code keeps a limit of its own.
    • Apps fit a phone. An app opened from its link now fits the screen under its notice, so the bottom of a game is never cut off.
    • Games take keys at once. On a desk, an app's own page gives it the keyboard as soon as it loads, so a game's arrow keys and Space work without a click first.
    • Agent builds count fairly. A build askr turns down because another is still running no longer uses up one of an agent's builds for the hour.

    Also

    • What it costs. Looking, opening and playing are free. An answer from askr AI inside an app is about $0.002 and never more than $0.05, paid by whoever is using it, after they say yes once. Putting something on Explore is $1, once. askr building an app is about 5 to 10 cents from the credits of whoever asks, and launching a page an AI wrote costs nothing. A deep build is paid by the step, about $0.09 each on the model that builds today, and never more in all than the most you set. Pictures, spoken lines and music made for an app are each priced before they are made, and engines, saves and worlds cost nothing to use. The first picture askr makes for something on Explore is free, and each one after is under 10 cents. On a sale the maker keeps 80% and askr keeps the rest, less the royalty where what was sold is a copy. askr's fee is in what AI usage costs and is never added to a purchase or to the $1.
    • The checks. askr's safety check reads every app before it is hosted and again before it goes on Explore, and reads every skill and framework for passwords, keys, seed phrases and wallet tricks. Apps run walled off, with no sign-in and no way to reach the internet: an app loads only its own pictures and sound, and an engine from askr's own shelf. Three reports take something off Explore until askr's team has looked.
    • Help. Explore has finding and opening things, Put yours on Explore has listing, the remixing choice, pictures and screenshots, Prices on Explore has paying and which credits can buy, and there are articles on communities, reviews and comments, skills and frameworks, your maker page, sales, earnings and royalties and AI agents on Explore. Chev has running a framework and paying its maker, and How credits work has the fee. Build something big has deep builds, an app's pictures and sound, engines, saves and worlds, and Write your own app has what an app's own code can call, the limits and what an app cannot do.
    • What the rest costs. A game, a group skill, a song, a clip or a voice room in a chat is paid like an @askr message, by whoever starts it. That is the chat's credit, or their own balance in a chat of their own and wherever each person pays for their own, which is where a public collab starts. Each says its price first, and askr's fee is inside it. Secret Santa, the ledger, a challenge, a savings circle, leagues, profiles, notifications, the daily game and drawing on a canvas cost nothing.
    • Help for askr with friends. Games in a chat, The Arcade, cups and leagues, Voice in a chat, Group skills, Teams and paid chats, Notifications, Profiles and leagues, Make it together, Videos together and Your look and voice are new. Chat with other people now has invite links, guests, polls, the price check, gifts and tips, and Pictures, clips and songs has giving something a page of its own.
    • Privacy and terms. Help's privacy page says what is public on Explore, and what askr keeps of your purchases and of what you open there, of your look, of voice rooms and voice answers, of your profile, and of your saves in apps and games. The privacy notice says the same of Explore, your profile, your look and your voice, and which suppliers make the clips, voices, songs, games and builds in a chat that nobody picks a model for, and hear its voice clips. It also now covers paying with any coin, what WalletConnect's connect window reports, and an AI agent paying in USDC. The terms gain a clause on buying and selling on Explore and rules for an AI agent paying in USDC, and their referrals clause now names the gift for a friend's first top-up.

    Prepped for what is next

    • Your own voice. Read a phrase askr shows you, and askr can read in your voice for the people you give a pass, chat by chat. You are told every time, and you can delete it and anything made with it.
    • Sound effects in apps and games. A jump, a door or an explosion, made by askr from a few words as it builds, beside the pictures, spoken lines and music an app already gets.
    • Clips with you in them. After a one-time check that your look is you, friends you allow can put you in clips. Pets and characters come with it.
    • A stream's own chat. Streamer mode reading the lines that start with !askr in a stream's chat on Kick and YouTube, into the streamer's queue.

    Finished: Fees on (and still lower than everyone else)Finished: askr RoomsFinished: Room walletFinished: Group skillsFinished: askr StudioFinished: askr PlayFinished: askr PagesFinished: Explore

    Announcement to follow.
17. 1.84Interim improvements8 October 2026Smaller changes between the big releases, added as they land.

    New features

    Nothing this time.

    Improvements

    • Chev steps aside while you chat. On a phone, once a conversation is under way, Chev sits small at the top left with its name and what it is doing beside it, so your messages get the screen. The big Chev still greets a new conversation.
    • Apps built in a chat cost less. A change now rewrites only the files it changes, and askr reads the newest copy of each file rather than every copy before it, so a message on a growing app costs a fraction of what it did.
    • Your app keeps its builder. Once a chat is building an app or a site, the model writing it keeps writing it, even when you push back or ask for something hard, so the next version is never someone else's rewrite.
    • Change the app you just made. In a chat of your own, asking for a change to an app askr built, like "make the questions harder", updates that app. No @askr needed.
    • Chev builds where your app is. A change you ask Chev for carries on in the chat that holds the app's code, a build that stops early is finished there, and each scheduled job works in a fresh thread.

    Bug fixes

    • A public chat runs on its own credit. In a public collab askr now spends only the chat's own credit, so nobody who joins can reach its creator's balance. The wallet says why the creator's top-up switch is off there.
    • One bad app file can't stall askr. The safety check reads an app's file once and stops any check that runs past two seconds, so a hostile file can no longer hold everyone else up.
    • Continue carries on where it stopped. A cut-off answer is finished by the same model, and a long file picks up where it stopped instead of starting over. Hosting an app again under the same name now updates it, on the same link.
    • A cut answer says why. When your balance covered only a shorter answer, the line after it says so and what it cost, and a shortage on askr's side never cuts code short.
    • Apps that save things run in the Preview. To-do lists and trackers no longer stop on their first line in the Preview or on a hosted link.
    • App builds finish. A build no longer runs out of room thinking on Claude, and when one can't finish it says so plainly and nothing is charged.
    • One press, one build. Pressing Enter twice on New app no longer builds or charges twice.
    • Chev says when it couldn't finish. Chev no longer closes a job as done when a step came back empty, and an answer with nothing in it costs nothing.
    • A huge post can't stall askr. A verdict's words, a chat's code and a builder's answer are each read once, so one oversized post can no longer hold up everyone else.
    • A picture that isn't kept costs nothing. A picture in a chat is charged only once it is saved; one that came back but could not be kept is free.
    • An app is charged once it is saved. askr saves an app's new version before it charges for it, and never over a newer one.
    • The privacy notice names everyone. It lists every company askr sends anything to, including the ones that answer for askr Pro and Chev.

    Announcement to follow.
18. 1.83Rooms with friends7 October 2026Bring anyone into a chat with one link, let the chat pay, build a small app together with "@askr build us a …", and give anything you make its own public page.

    New features

    • Invite links and QR codes. Every chat has a link to share and a QR code for someone across the table, and anyone in the chat can invite unless the person who started it says no.
    • Join with just a name. A friend who opens the link joins as a guest with only a name: no sign-up, no email.
    • Report, block and mute. Report any line, block someone so their lines are hidden for you everywhere, or mute a chat so it stops counting as unread.
    • Side rooms. Take this to a side room starts a new chat with the people in the thread, linked to the first.
    • Catch me up. "@askr catch me up" sums up, for you alone, what happened since your last message, every quote linked to its line. Newcomers get a short welcome of their own.
    • Ask privately. Ask askr something inside a chat and only you see the answer.
    • Polls. Two to ten options, a timer and results hidden until it ends, or Make a poll from this chat and askr lists what people put forward.
    • Room modes. One tap makes askr the chat's Referee, Quizmaster, Planner, Translator or Note-taker.
    • Read the chat in your language. Other people's messages are shown to you in the language you pick, marked Translated, with the original a tap away.
    • askr can speak up, if you let it. The chat's creator can let askr speak on four moments only: a question nobody answered, a decision that stalled, something that looks wrong and dates that clash. Anyone can mute it.
    • The chat keeps answering. When a line needs more than the chat's credit holds, the creator's balance tops it up by exactly what is missing.
    • Who used what. An itemised bill in the chat's sheet: every line that cost something, who asked for it, on which model, what it cost and which balance paid.
    • The price before a costly line. A clip, a song or a big site says its price under the line and waits for Run it.
    • Who pays, and limits. The creator chooses who pays (the chat, each person, or the chat up to a daily limit and then the person), sets daily limits per person if they want them, and rules for a team's chat.
    • Chip in, gift and top up. Split a costly line with friends, give the chat 20 answers, $5 or $10 as a present, or top a chat up by card or crypto.
    • Tips. Tip a great answer or creation $0.25, $0.50 or $1, from credits you did not buy by card.
    • A page for anything you make. Publish a page from any answer, picture, song or clip at heyaskr.ai/s/…, with the recipe, Make my version and votes, while the chat stays private.
    • Daily challenges in askr Community. A picture theme every day, a song of the day and a video theme every week, with small prizes in askr credit.
    • Verdicts, the gallery and the askr mark. Settle this verdicts get pages of their own, the best pages are at heyaskr.ai/gallery, and downloads carry "made in askr by @you".
    • "@askr build us a …". askr makes a one-file app for everyone in the chat (a tracker, a poll, a leaderboard, a bill splitter, a game), priced before you send, about 10 to 15 cents.
    • Apps the whole chat shares. The app is pinned in the chat's Apps strip, and one person's tick, vote or move shows for everyone at once.
    • Change it in words or in code. "@askr make the scores bigger" makes version 2, or edit the code yourself and save it.
    • Links, Remix and a gallery of apps. Every app opens from a link without signing in, Remix copies one into your own chat, and heyaskr.ai/apps shows the most used each week.
    • Starter apps, build battles and a weekly jam. 18 starter apps in askr Community, three models building one prompt side by side, and a weekly theme with prizes in askr credit.

    Improvements

    • Apps run walled off from askr: they never see your account or your session, cannot reach the internet, and open a web page only after you say yes. askr's safety check reads every app before it is kept.
    • A build never costs more than the most shown with its price, one that does not come costs nothing, and using an app is free.
    • Someone you block cannot invite you or add you to a chat.

    Bug fixes

    Nothing this time.

    Finished: askr BuildMoved on: askr RoomsMoved on: Room walletMoved on: askr Pages

    Announcement to follow.
19. 1.82Chev6 October 2026Meet Chev, your own agent on askr.

    Text it a job, in the app or on Telegram, and it opens a collab, asks models there while you watch, and comes back with the answer; what it cost is a tap away, on the job's card and in Chev settings. It runs jobs on a schedule, reaches you where you are, works in the collabs you share, and can hold a wallet of its own. Chev does the doing itself and hands real thinking to askr, at the level you set, and it spends your credits inside the limits you choose.

    New features

    • Meet Chev. Chev joins the menu for every signed-in account, between Chats and Library. Meeting it walks through its set-up, one screen each, with Back and Next: a name (your @ plus Chev to start) and a look (twelve colours and fifteen accessories), how it talks, how it works for you (how hard it thinks, when it checks with you, and a daily limit if you want one), where it reaches you (Telegram and notifications, which can both wait), and a first job from a few starters. Every choice starts on a default and is saved as you make it, and Skip the rest opens the chat at once. Until it is finished Chev opens on the same screen again, after a reload or on another phone or computer, and Set up again at the foot of Chev settings walks through it once more.
    • Text it a job. Its conversation at heyaskr.ai/chat/chev takes a message like a text to a friend, with up to 20 photos and documents or a voice note of up to two minutes. Over the box, jobs that start in one tap: Watch a price for me, Every morning the news on a topic, A weekly summary of my collabs, Track a launch date, Remind me of something.
    • It works in collabs, while you watch. For a real job Chev opens a collab of its own, asks models there as itself, compares and checks their answers, posts a report and finishes with a short message. Its name at the top of the chat says what it is doing (Ready, Typing, Waiting for you, In line, Done). While a job runs, its card under your message says what it is doing now, with Watch to open the collab over the chat and Stop to stop the job. A plain reply is just a message; a job that did work keeps one small card that says what it did and opens on a tap: the collabs it opened, Chev's lines about it, and what it cost. Links in an answer are short: another site shows by its name, and askr's own pages are small buttons that open in the app. A collab Chev opened stays out of your chats list until you write in it, share it or invite someone; until then it opens from the job's card and from Its work.
    • Go deeper, Another angle, Make it shorter. Under a finished job, in the chat and as buttons on Telegram and Discord, each carries the job on in the same collabs.
    • A doer, with askr to think. Chev does things itself, quickly: it replies, sets the reminder or the watch, sends the message, starts the work. When a job needs real thinking (research, comparing, a plan, more than a few lines) it hands the whole question to askr in a collab and brings the answer back. How hard it thinks is Cheap, Balanced or Performance, under Limits, Balanced to start: that is how hard askr thinks when Chev asks it, with askr Pro picking the model. Chev's own small steps cost the same at every level. No model is named anywhere in Chev.
    • A job that keeps going. Up to 16 steps, 8 model asks and 30 minutes; a long job of yours carries on in up to four parts, in up to four collabs. Chev asks you a question only at a real fork, before spending a lot, or when stuck, with options as buttons; a question left for a day closes on its own. While it works, "stop", "cancel" or "halt" stops it, "wait" or "hold on" holds it until you say what to change, and "status" says what it is doing. If your credits run low a job of yours pauses, says so once, and carries on by itself when you add credits. Five jobs on at once; a sixth is turned down until one is done.
    • Through askr's own updates. A job carries on across a deploy: its steps are saved as they happen, a step cut off in the middle is picked up where it was, never run twice and never charged twice, and the model it thinks with stays the same for the whole job. Two processes never run one job.
    • What it can do. Search the web and read links and files; make pictures, clips and sounds; build a page or a small app in the collab's code panel; load one of askr's skills; keep to-dos, notes and drafts; set reminders (free: no model is involved); email you something you asked for, up to five a day; remember a fact about you in askr Mind; invite people to a collab, keep notes for a collab on a schedule, digest your collabs, and ask another person's Chev. It can change its own look or name when you ask. Before it invites anyone to a collab or makes a room public it asks you, every time.
    • Videos: read first, watch when asked. Sent a video, or a link to one on YouTube, TikTok, Instagram or Facebook, Chev reads what is said in it. It watches the video only when the question is about what is on screen, since watching costs more; where it cannot watch, it reads and says so plainly.
    • Prices and price watches. Ask what a token is worth and Chev reads it live, with no search: the price, market cap, liquidity and the day's move, for $ASKR, the big coins, or any token by its ticker, its contract or a link to it. Ask to be told when a price or a market cap reaches a level and it sets a watch: checked every minute, free since no model is involved, and you are told once when it gets there. Each watch is under Its work, Watching, with the figure now and Stop; up to 20 at a time.
    • Schedules and reminders, on your clock. Every day, on weekdays, every week, every few hours, or once, at the time you name in your own time zone, which Chev learns from your browser. A scheduled job keeps it brief and stays quiet when there is nothing new. Each is listed under Its work, Schedules, with when it next runs and what it reported last time, to pause, change or delete, and New schedule there sets one up without asking Chev. Every Chev has a Weekly review on Sundays at 18:00 UTC: what it did and up to three schedules it suggests; a week with no work is skipped.
    • Its own Telegram bot. Connect Telegram in Chev settings or when you meet it, tap Start, and text it there: voice notes of up to five minutes, photos, documents and albums. One progress message per job is edited as the work moves on, with Watch and Stop, and goes away when the job ends, so the chat reads as your message and the answer; the answer carries Open the collab, Full report here, Go deeper and Another angle. Only the newest message keeps its buttons: a question's go when you answer it, the last answer's when the next job starts. Voice replies on Telegram adds a short voice note saying the same as each reply. Private chats only. This is Chev's own bot; the askr community group's bot is another thing.
    • Notifications on your phone or computer. Under Connections, Turn on asks your browser, and Chev then tells you on that device when a job is done or it needs an answer, even with askr closed, though not while you are in its chat. On an iPhone or iPad, from askr on your home screen.
    • Where its own messages land. Chev answers where you asked: a job from the app answers in the app only, a job from Telegram or another channel answers there, and every line shows in the conversation in the app whatever channel it came from. Each event tells you once, by the channel it is on or by a notification from the app, never both. Reminders and scheduled jobs picks the one place for what Chev starts itself: the app, Telegram, or a channel you connected. Ask it to message you on Telegram and it sends that one message there.
    • Chev in your collabs. Write @chev in a collab you are in and Chev works there, in front of everyone. Each member's Chev is in a strip over the message box, under its own name, with Thinking while it works; a tap puts its @ in the box. @chevs asks every Chev in the collab, up to six, each on its owner's credits.
    • Other people, only if you say so. Let other people ask my Chev is off until you turn it on. Off, when someone names your Chev it asks you first: Answer for me or Ignore. On, anyone in a collab with you can ask it by name, it answers there on your credits, up to ten asks a person a day and three model asks a job, and its answer says whose credits paid. Work for someone else sees only that collab and what you let it share, never your memory, notes, schedules or other collabs.
    • Join in, stand in, catch me up. Join in when a room allows it lets your Chev speak up unasked where the collab's creator has turned on Let Chevs join in, on an unanswered question, a wrong fact, a plan or a decision worth writing up: at most one Chev line for every eight from people, three an hour and ten a day. Stand in for me when I'm away answers for you after half an hour away, from the collab and What it may share about you, marked as your Chev. Catch me up, offered once more than 15 lines have come in since you last looked, has your Chev sum up what you missed in your conversation with it, never in the collab. Decided here, in the collab's sheet, keeps what the Chevs in it noted as decided.
    • Your credits, your limits. Chev spends your balance through the normal ledger, priced as if you had asked yourself, and askr adds nothing for Chev. No price shows in the chat until you tap a finished job's card; each job's cost is also under Its work, Recent work, and Spent today is under Limits. Set a daily limit is off until you turn it on, then any amount from $0.01 to $10,000: at the limit a job of yours waits and carries on when you raise it, anything else stops for the day and says so. When it checks with you: Ask me first, Do it and tell me (where every Chev starts, asking first above $1, an amount you can change), or Just do it. Pause stops anything new until Resume.
    • A wallet for Chev. Under Chev settings, Wallet, Give Chev a wallet: one wallet for each Chev, made and held on askr's server with Privy, on Base, Ethereum, BNB Chain, Robinhood Chain and Solana. Add money by sending a dollar coin from your own wallet to its address; it shows a minute or so after it lands and Chev tells you. askr pays the network fee on every chain for up to twenty transfers a day (three on Ethereum), each of at least 10 cents ($1 on Ethereum); past that a transfer goes the next day, or uses the chain's own coin if the wallet holds some. A wallet holds dollar coins only: SOL sent to it by mistake shows on the tab with Send it back, and ETH or BNB show with a note that the key can move them.
    • How Chev pays. A new wallet starts on Ask me first: nothing is paid without your yes, which you give in the chat, on Telegram with Approve and Decline, or on the Wallet tab, and which lapses after twelve hours. Up to an amount pays by itself at or under your amount; Just do it pays and tells you after. Two limits, Most in one payment and Most in a day, hold in every mode, even against your yes. Chev's pay tool sends a dollar coin to an address or pays a seller that asks for payment the standard way (HTTP 402) and gets the seller's answer back; the same payment is never sent twice.
    • Your money stays yours. Take money out sends everything, or an amount, to your own address at any time, never limited. From your own wallet lets Chev draw up to an amount straight from your own wallet, without moving the money first: one approval in your wallet, the coin's own contract holds the cap, 0 ends it. Take the key out opens the wallet's key in your browser only, encrypted on the way, so askr never sees it; whoever holds the key can move the money without Chev's limits, and the tab says so first.
    • Its settings, in five tabs. Personality: its name, look, how it talks (Straight, Warm, Playful, Formal), What it should know, and What Chev knows about you (your askr Mind, to reword or forget in place). Limits: the daily limit, when it checks with you, how hard it thinks. Wallet, where this askr has wallets open. Connections: Telegram, notifications, where reminders land, and In your rooms; more channels show there as they open. Its work: Schedules, Watching, To-dos, notes and drafts, and Recent work, its newest ten jobs with Show all.

    Improvements

    • A collab knows its Chevs. The sheet of a shared collab gains Chevs here, the creator's Let Chevs join in switch (off by default), and Decided here. The @ picker lists each member's Chev by name, with @chev for your own and @chevs for all.
    • askr Mind, from Chev. A fact you tell Chev worth keeping goes into your askr Mind, so every model you use knows it next time, and shows in Settings, Memory like any other note. Chev reads your memory before every job, and is told none of it while askr Mind is off.
    • Only your messages instruct it. What Chev reads in collabs, model answers, web pages and files reaches it marked as information to weigh, never as instructions, whatever the text says. It never reads or writes in an askr Anon chat.

    Bug fixes

    • HEIC photos on Android. A photo saved as HEIC now attaches from Chrome, Firefox and Edge, on a phone or a computer, where it could not be read before; it already did on an iPhone and in Safari. It is turned into a JPEG on your device before it is sent, in a chat, a collab, with Chev, in Community and as your profile picture. A HEIC sent to Chev as a file on Telegram is read as a photo too, where Chev used to take it for a video.

    Also

    • What it costs. Nothing for Chev itself. What Chev does itself (a reply, a reminder, a watch, sending a message) costs well under a cent. A question it hands to askr costs what the same question would in a chat of yours at your level, a few cents on Balanced. A picture is about $0.04 and a clip $1 to $3. Reminders and price watches are free, and failures are not charged.
    • Help. Chev has every step: meeting it, giving it a job, what it can do, schedules, each channel, Chev in your collabs, the three levels, money and the wallet, stopping, and what it remembers.
    • Privacy. Help's privacy page says what askr keeps of Chev: your conversation, each job's steps and cost, its schedules and lists, what you connected it to, and its wallet's payments. Remove Chev deletes its jobs, schedules, lists and connections; its collabs stay yours.

    Prepped for what is next

    • Chev on Discord, by email and on iMessage. Direct messages on Discord, an email address for each Chev that only its owner may write to, and one iMessage line to text. Each switches on as its account opens, and Connections lists them as they do.
    • Cards for Chev. Your own Visa through Link, with a yes on your phone for each purchase; a Visa on Chev's wallet that spends its dollar coins; or a card you give it. Every purchase under the wallet's modes and ceilings.
    • Chev buys things. Chev opening a shop's page and filling the checkout itself, which is what makes a card useful. Until then it pays sellers that take payment on the open x402 standard from its wallet.
    • Chev runs frameworks from Explore. On a job or on a schedule from a framework's page, as a real run, paying the maker's share from its wallet, asking you first unless you say otherwise. Lands with Explore.

    Finished: Chev

    Announcement to follow.
20. 1.8askr Pro improvements2 October 2026askr Pro now owns the result and learns which models work.

    Ask for the whole piece and you get the whole piece. When a model declines, fails or costs more than your balance, askr Pro moves on by itself instead of asking you to pick. A model checks every answer, what you do next counts too, and once the record is clear askr Pro picks by what has worked, everyone's first and then yours. Today's other fixes ride with it: askr makes PDFs you can save, and every message can search the web when it helps.

    New features

    • The whole piece, written. Ask for an essay, a report or a piece of coursework and askr writes all of it, in natural prose at your level and in your language, not a plan or an outline. Sources it cannot know are left as [placeholders] for you to fill in, never made up. The Learning skill writes it too when you ask for the finished piece, then offers to walk you through it.
    • Past a model that declines. When a model declines what you asked, or you push back on its answer ("no, just write it", "that's not what I asked", in any of seven languages), your next message goes to another company's model at the same level. Details says "in place of" the model it moved on from.
    • Skip a maker by saying so. Say "don't use Claude" and askr Pro leaves that company's models out of its picks for you, in English, Dutch, German, Spanish, French, Portuguese or Italian. askr Mind keeps it as a note you can see. Say "use Claude again", or forget the note in Settings, Memory, to have them back.
    • Failures recovered. When a model cannot answer just now, askr Pro tries once more by itself, on the same model through its maker's own connection or on the next model at your level, and the chat says so. The first try is not charged.
    • A lighter model for a short balance. When your balance cannot cover the model askr Pro picked, it answers with a lighter one your balance does cover, and says so. The same goes for pictures, clips and sounds. Out of credits shows only when nothing fits.
    • A check on every answer. A low-cost model reads each askr Pro answer beside the message that asked for it, and says whether it did what was asked, did part of it, declined or missed. askr keeps the verdict, never the words.
    • What you do next counts. Asking the same thing again, pushing back, Try again, choosing a model yourself for the next message or tapping Stop count against an answer. Carrying on counts for it.
    • Picks by what has worked. It all adds up to a record for each kind of job and each model, at each level. Once the record is clear, never on one good or bad answer, askr Pro puts the model that has worked best first. On Cheap, a model that works as well for clearly less goes first. It starts from everyone's record, and once you have enough answers of your own for a kind of job, your own counts for you. Details under the answer says "what has worked best".
    • Trying a model now and then. About one message in twenty that starts something new goes to another model at the same level, so the record keeps up as models change. Details says "trying it out". It never does this on Try again, or right after you push back.
    • A second answer on Cheap. When the check finds a Cheap answer fell short, a stronger model of another company answers too, right under it, at its own model's price. Details says what happened, such as "askr Pro: Gemini 3.8 Flash fell short, Claude Sonnet 5.5 answered too".
    • PDFs, made for you. Ask for a PDF, or for a document to save, print or send, such as a CV, a letter, a report or an invoice, and askr writes it as a document in the chat with Save PDF, A4 with page numbers. askr Pro sends a PDF ask to a model that writes, never to a picture or to code, in English, German, Spanish, French, Portuguese, Chinese and Japanese.
    • The web, whenever it helps. Every Ask and Code message can now search the web unless you turn the globe off, and the model decides when a message needs it. A model never tells you it cannot go online.

    Improvements

    • A model you choose lasts one message. Under askr Pro, a model picked in Choose a model, or with Use in this chat on the Models page, runs your next message, shown above the box with this message, then askr Pro chooses again. Tap Off in the askr Pro row to stay on a model.
    • Try again asks another model. Under askr Pro, Try again on any answer goes to another model at the same level, not the one that just answered. On a message you sent with a model you chose, it uses that model again.
    • Every pick says what happened. Details under an answer now says when askr Pro moved "in place of" a model, when one "answered, X could not", "a lighter model for your balance", "what has worked best" or "trying it out".
    • Failures in askr Pro's words. When nothing could answer, the chat says which models askr Pro tried and that nothing was charged, and never asks you to pick another model.
    • Attachments a model can read. askr Pro no longer picks a model that cannot see what you attached, such as a picture or a clip.
    • Long clip and sound requests. A request too long for the clip or sound model is shortened once and run again, instead of failing.
    • Your latest message decides. Every model is told to deliver what your latest message asks for, the whole piece, the length or the format, and that your preferences in memory come before a skill's habits.
    • Each model speaks for itself. A model no longer apologises for, or speaks as the author of, an answer another model wrote, and never presents a skill's own rules as rules for every model on askr.
    • Standing wishes remembered. askr Mind keeps the "next time", "from now on", "always" and "never" wishes you give about how askr behaves, in your own memory across chats.
    • Looking things up under askr Pro. When the model askr Pro picked cannot search the web, it says the answer can be looked up if you ask, instead of telling you to switch.
    • Use in this chat. Opened from a chat, the Models page puts the model you choose into that chat and the conversation carries on. Use this model says it starts a new chat.
    • A search your balance cannot cover. When your balance covers the answer but not a web search, the message runs without the search instead of being refused, and the model says a top-up lets it look things up.
    • How the buyback and burn are worked out. On the token page, each month's estimated buyback and burn counts its tokens at today's price, How do we calculate this? under each estimate walks through the sum in today's figures, and the week ring shows buybacks and credit airdrops as dots of their own.

    Bug fixes

    • A plan in place of the piece you asked for. Asked for a report or an essay, the writing skill could decline to write it and give a plan or an outline instead, again on every retry. It now writes the piece.
    • Asking again brought back the same model. Pushing back, or sending your message again, went straight back to the model that had just declined or failed. That model is no longer kept for the next message.
    • Try again on an answer chose the same model. Under askr Pro, Try again on an answer that ran fine came back on the very model you were trying to get past.
    • Out of credits asked you to pick a cheaper model. Under askr Pro, where you pick no model, a short balance now gets a lighter one instead.
    • A moment on askr's side read as the model's refusal. When askr could not reach its models for a while, the chat said the model's provider had turned the request down. It now says it is on askr's side, and that nothing was charged.
    • A model said models cannot be skipped. Asked not to use one, a model could say that models cannot be chosen or turned off in askr. Every model under askr Pro now knows they can.
    • The Learning skill told you to hand in your own work. Its hint panel no longer says so.
    • A model said it could not browse. Asked about a link or something current, Claude, GPT and askr Pro could answer that they had no browsing tool. They now search, unless the globe is off.
    • A chat others wrote in could not be ended. Once everyone else had left a Collab, its creator saw only Delete, which askr refuses for a chat other people wrote in. It now offers End Collab, which ends it, returns what nobody spent and keeps it readable.
    • Token estimates jumped during the day. The month's estimated buyback and burn counted today as a whole day, so the first buy of a day halved the pace. They now count ended days only and move once a day, at midnight UTC.

    Also

    • Built from the models that exist. askr trains no model of its own. askr Pro is built from the models that exist, arranged around you: it learns which of them to pick for you.
    • What it costs. The check is free, like askr Pro's choosing. A second answer on Cheap is charged at its model's price and shown under it, as on every answer. A step down for your balance costs the lighter model's price.
    • Never in askr Anon. Nothing in an askr Anon chat is checked or counted.
    • Help. askr Pro says how it learns, the second answer on Cheap, how to skip a maker, and what it does when a model declines, fails or costs more than your balance.
    • Privacy. The privacy notice and Help's privacy page say that a model checks each askr Pro answer, never in askr Anon, and that askr keeps the verdict, not the words.

    Moved on: askr Pro

    Announcement to follow.
21. 1.78askr Anon2 October 2026Private chats, when you need them.

    askr Anon needs no account, locks every chat with its own key made in your browser, answers on sealed models first, and erases everything when the chat ends.

    New features

    • No account. Open Anon in the menu or go to heyaskr.ai/anon, and tap Start a chat. askr gives your browser a random guest name and a sign-in cookie, and nothing else: no email, no password, no wallet. Tap Keep this account to add a password, still with no email.
    • Locked with its own key. Each chat gets its own key pair, made in your browser (ECDH P-256). The private half stays in your browser and in the invite link, after the #, the part of a link browsers never send to a server. askr gets the public half only and stores every message sealed to it (AES-256-GCM), so the database holds scrambled text that askr cannot read back.
    • Held only while it answers. askr cannot open a chat's history, so your browser opens it and sends it with each message. askr holds the words in memory while the model answers and writes down only the sealed copy. Server logs keep no words.
    • Sealed models first. askr Pro starts on a sealed model: DeepSeek V4.1 Flash on Cheap, GLM-5.3 on Balanced and Kimi K3 on Performance, stepping to GLM-5.3 Flash, GPT-OSS 120B and the others in its tier if one is down. Gemma 4 31B and Llama 3.3 70B are sealed too, to pick by hand. Sealed models are open models running inside secure hardware that proves the exact code it runs before askr sends a word, so once your words leave askr nobody can read them, not the host and not the gateway.
    • Zero retention after that. If no sealed model in your tier is answering, Pro moves to zero retention endpoints of models such as GLM-5.3, Kimi K3 and DeepSeek V4.1 Flash, whose companies keep nothing once they answer, by their own policy. 147 models carry the zero retention mark.
    • Routing that never reads your words. In askr Anon no decision model reads your message to choose a model. Pro picks by code alone, from the tier you set: Cheap, Balanced or Performance.
    • Any model, with a warning first. Pick any model yourself and the chat says who keeps your words and for how long before you send, such as "Anthropic keeps what you send for 30 days." Pictures, clips and sounds have no sealed or zero retention model yet, so you pick those yourself, with the same warning.
    • One random id per chat. A model company gets one random id for the whole chat for its abuse checks, keyed with a server secret, never your account, your name or your IP address.
    • Nothing else reads it. No askr Mind, no web search, no projects, no comparing models, and no helper model: no prompt writer, no captioner, no decision model. Nothing in it is indexed for recall.
    • Gone when it ends. End and erase deletes everything for everyone at once: the messages, answers, pictures, clips, sounds and files. A chat nobody writes in for a day erases itself. Credits nobody spent go back to whoever added them, and copies of the database kept for recovery age out within 16 days.
    • Invite by link, pay together. Anyone with the invite link joins with no sign-up, guest or not. Once more than one person is in, askr answers from the chat's own credit, which anyone in it can top up from their balance. Guests add credits with crypto, from hundreds of coins with Monero among them, or with any token from a connected wallet, never by card, because a card payment gives Stripe your name.

    Improvements

    • Every model says what its company keeps. Sealed, zero retention, or kept and for how long, wherever you pick a model.
    • Deleting means deleted. A deleted file takes its name and extracted text with it, a forgotten memory takes its words, and expired sign-in codes and sessions are cleared every hour.
    • Nothing to identify people in public chats. Someone watching a public chat no longer receives the account ids of the people in it, or who added credits.
    • A keyed id for abuse checks. The id some model companies get for their abuse checks is now keyed with a server secret, so nobody can work it back to an account.
    • No analytics. The page-view counter is gone from every page.
    • Zero retention routing. Models marked zero retention are now always sent to their zero retention endpoints.

    Bug fixes

    • Memory off means off. A chat with Remember this chat turned off no longer sends its lines to be indexed for recall.
    • Logs keep no words. Server logs no longer carry answers, refusals or memory entries, only their length and status.

    Also

    • Privacy. The privacy notice now says what is kept and for how long, cookie by cookie, and names every place an IP address is written down.
    • Help. askr Anon has every step, what askr Anon keeps and what it leaves out.
    • Where it stops today. The words pass through askr's server while a model answers. Pictures, clips and files are stored as they are until the chat is erased. askr's access log records IP addresses, rotated and never tied to an account. Paying from a connected wallet, or with a coin whose sender shows on chain, links that payment to the guest.
    • Next: end to end on sealed models. Your browser checks the sealed hardware's proof itself and locks each message straight to it, so askr only relays scrambled words. Answers come back locked to your browser and are stored sealed to the chat's key, so askr never holds a readable copy, not even in memory.
    • Next: files locked like messages. Pictures, clips, sounds and files stored sealed with the chat's key.
    • Next: no IP addresses kept. Anon requests leave no IP address in askr's logs, and askr gets a Tor address, so you can reach it without showing yours.
    • Next: payments that can't be tied to a chat. Zero-knowledge payment proofs, so askr can check a chat is paid for without knowing which deposit paid, and prepaid codes you can buy anywhere and redeem as a guest.
    • Next: check it yourself. Open code for everything that touches your words, a public record of exactly which build is running, askr's own relay inside sealed hardware with its proof published, and an outside security audit. The roadmap tracks each step.

    Moved on: askr Anon

    Announcement to follow.
22. 1.77Attach and animate2 October 2026Attach up to 100 pictures and 10 videos on a message, and turn up to ten pictures into one video: a clip from each, or one long move from picture to picture, with a song under it if you want one.

    Files attached in a chat reach the model again.

    New features

    • Up to 100 pictures on a message. It was 6. Drop in a whole set of screenshots and ask about all of them at once. If they will not fit the model you picked, askr says so before anything is charged.
    • Up to 10 videos on a message. It was 1. Each can be up to 100 MB and 10 minutes, for the models that watch video, such as Gemini.
    • One video from your pictures. Attach two to ten pictures and ask for a video. askr plans the shots from all of them together, makes a clip from each and joins them into one video, in the order you attached them.
    • Transitions. On Kling 3.0, Kling O3 and Veo 3.1, each clip can run from one picture to the next, so the video moves through your pictures as one.
    • A song under it. askr makes music the length of the video and lays it under the whole thing.

    Improvements

    • Priced in full first. The price beside Send is the whole video, every clip and the song, and credits for all of it are set aside before anything starts, so a video never stops halfway.
    • You pay for what is made. Each clip is charged when it lands. A clip that cannot be made is tried once more, then left out and not charged.
    • Progress on the line. While the video is made, its line says how many clips are done, for everyone in the chat.
    • Under askr Pro, just ask. "A 30 second promo from these, with transitions and music" sets the length, the style and the song.
    • Big batches always send. A message's files travel inside the request up to 8 MB, and the rest go to the model by a private link that expires within the hour, so a large batch is never too big to send.
    • A batch is not paid for on every reply. More than six pictures, or more than one video, on one message go with it and with your next message, then askr names them instead of sending them again. A few pictures or one video carry on with the chat as before.
    • Up to 20 files on a collab message. It was 6. The room pays for what its members attach, so a collab keeps a count.
    • Smoother uploads. Pictures are resized and sent three at a time, so a phone stays responsive with a big batch, and the files above the box scroll after three rows.

    Bug fixes

    • Files in a chat were not reaching the model. From 26 September, a picture, PDF, document, audio file or video attached in a chat was named to the model rather than sent, so the model could not see or read it. Collabs were not affected.
    • A chat stopped taking pictures after 20. Pictures from earlier in the chat counted even though they were no longer sent. Now only what is sent counts.
    • A collab's Video door refused pictures when the collab's chat model could not see them. A picture there starts the clip, so the clip model decides.

    Also

    • Help. Photos and files lists what each kind takes, and One video from several pictures has every step and what it costs.

    Announcement to follow.
23. 1.76askr Swap1 October 2026Pay for credits with any token on Robinhood Chain.

    Your own wallet sells it and the dollars reach askr in the same transaction, and you see what you get, and every cost, before you sign.

    New features

    • Tokens in Add credits. Under Crypto, choose Majors (BTC, ETH, SOL, USDC and more) or Tokens: ETH, USDG, $ASKR, any Robinhood stock token, or any other coin with a market on Robinhood Chain. Your wallet sells the coin for USDG, a dollar coin, straight to askr in one transaction, so askr never holds your coin.
    • Your wallet, by name. The wallets in your browser are listed with their own names and icons, such as MetaMask or Rabby, and WalletConnect opens a wallet app such as Robinhood Wallet. Connecting asks for no signature.
    • Your tokens, ready to pay. Once you connect, the tokens askr can price show straight away, most valuable first, with their logos and what each is worth in credit. Search finds anything else by ticker, name or contract address, even a coin launched today.
    • See it before you pay. askr test-sells your amount first. The quote leads with what you get and the least you can get, then a safety check in green, amber or red. Price details lists the coin's own tax, pool fees, price impact, askr's fee and the network fee.

    Improvements

    • Credits in seconds. The credits land once the sale is confirmed on Robinhood Chain, usually within seconds. The sheet says what was added and with which coin, for example "Added $21.03 with PONS", and so does the receipt.
    • Every top-up by name. Activity names the coin of every crypto top-up, such as "Added with BTC" or "Added with PONS", and your $ASKR bonus has its own row.
    • Your bonus applies here too. A top-up with tokens gets your $ASKR deposit bonus of up to 25%, like any other deposit, and whoever invited you gets their 10%.
    • Add credits remembers. It opens on Majors or Tokens, whichever you used last, and Pay from wallet under Majors lists your wallets by name too.

    Bug fixes

    • Pay from wallet picked a wallet for you. With two wallet extensions installed, it used whichever one had taken over the page. Now you choose.
    • A deposit bonus read "Credit added" in Activity. It now reads "Deposit bonus".

    Also

    • What it costs. askr's fee is 1% of the dollars the sale brings in, at least $0.05, and the quote shows it before you pay. A payment credits from $2 to $10,000. Card and Majors are unchanged: every dollar is a dollar of credit, with no askr fee.
    • Help. Pay with any coin has every step, cost and check.

    Finished: askr SwapMoved on: Our own payment processor

    Announcement to follow.
24. 1.75Token flywheel1 October 2026askr now buys $ASKR on the market every day from one public wallet and burns 75% of it every Monday, live on the rebuilt token page, and $ASKR holders get a bonus of up to 25% on every deposit.

    It takes the place of the temporary 95% holder discount.

    New features

    • The Flywheel. The token page tracks every buyback and burn as it lands, read from the chain: what askr has bought back and burnt so far, this month's estimates, the countdown to Monday's burn, every dev buy with its transaction, and the free credit airdrops.
    • A bonus on every deposit for holders. Hold $ASKR in a linked wallet and every deposit gets a bonus in credits of up to 25%: your share of supply × 100, so 0.1% of supply adds 10% and 0.25% or more adds 25%. Your holding is read once a day at 00:00 UTC, and a deposit counts the lower of that reading and what the wallet holds when you deposit.
    • One chart for $ASKR. The price since launch, with every burn, buyback and credit airdrop as a dot on the line where it landed. Point at one, or tap it, to read it.

    Improvements

    • The token page, rebuilt. Price, market cap, volume and holders up top; circulating, burnt and max supply in the details; what holding does under the Flywheel; the supply breakdown, the company's lock, the community and the proposals that shaped it below.
    • A discount on askr's fee. Your share of supply × 100 comes off askr's fee from the day it switches on: 0.1% pays 10% less of it, 0.5% pays half, 1% pays none. askr charges no fee yet.
    • Add credits shows your bonus. What you get, by card or by crypto, has the bonus folded in, with your own rate when a linked wallet lifts it.
    • Bonus credits stay with you. They are spent first, before the credits you paid for, and can never be sent to anyone; a room you fund gives them back as bonus credits.
    • Pricing, Help and Settings say what holding does now. List prices only on Pricing, the deposit bonus and the fee rule in Help, and your own bonus in Settings.
    • Roadmap in the bar. Roadmap sits between Token and Help at the top of every page.

    Bug fixes

    Nothing this time.

    Also

    • The temporary 95% holder discount ended with this release, as the buybacks began. Every change to $ASKR from here comes through a proposal, like Token Flywheel Proposals #1 and #2.

    Finished: Token flywheel launch

    Announcement to follow.
25. 1.74Use AI together30 September 2026The home page now leads with what only askr does: AI with friends.

    Group chats play on the first screen, Start a chat with friends starts one in a tap, and an invite-only chat now holds up to 100 people, the same as a public one.

    New features

    • Use AI together. The first screen of the home page shows group chats as they happen: friends talk, one of them asks @askr, askr Pro picks the model, and the plan, picture, song, clip, site, quiz, split or script lands in the chat with what it cost. The eight chats were made for real on askr and play on a loop.
    • Start a chat with friends. Under the Ask anything box, one tap makes a new chat and opens it ready to invite people by their @ or to share its link. Signed out, sign-in comes first and says why, and the chat is made as soon as you are in.
    • Chats of 100. An invite-only chat now holds up to 100 people, the same as a public one, you and invites waiting for an answer included.

    Improvements

    • How it works, together first. Four steps on the home page say how a chat with other people works, from starting one to how its credit works, and the questions answer whether your friends need to pay, how many people fit in a chat and whether askr works on your own.
    • No pile of empty chats. Start a chat with friends and Invite open the chat you made last and never used, on any device, instead of making another.

    Bug fixes

    Nothing this time.

    Announcement to follow.
26. 1.73Direct Seedance routing30 September 2026Seedance now runs on ByteDance's own API too.

    Pick the version marked (direct) under Seedance 2.0, Seedance 2.0 Fast or Seedance 2.5 and the same clip costs less than half: a 5-second Seedance 2.0 clip is $0.76 instead of $1.74. The versions you use today stay where they were.

    New features

    • Seedance (direct). Seedance 2.0, Seedance 2.0 Fast and Seedance 2.5 straight from ByteDance, listed under each model as its (direct) version, in the Video door and in rooms.
    • Less than half the price, at every length. A (direct) clip costs what ByteDance charges: 56% under the version beside it on Seedance 2.0 and 2.0 Fast, and 57% under on Seedance 2.5. The prices below are 720p in 16:9; 1:1 and 9:16 cost the same, and 4:3, 3:4 and 21:9, which hold a few more pixels, under 1% more.
    • Seedance 2.0 (direct), 56% under. 4 seconds $0.61 instead of $1.39, 5 seconds $0.76 instead of $1.74, 8 seconds $1.22 instead of $2.78, 10 seconds $1.52 instead of $3.47, 15 seconds $2.27 instead of $5.22.
    • Seedance 2.0 Fast (direct), 56% under. 4 seconds $0.49 instead of $1.12, 5 seconds $0.61 instead of $1.39, 8 seconds $0.97 instead of $2.23, 10 seconds $1.21 instead of $2.78, 15 seconds $1.82 instead of $4.17.
    • Seedance 2.5 (direct), 57% under. Any whole length from 4 to 30 seconds, from $0.93 instead of $2.18 for 4 seconds to $6.94 instead of $16.32 for 30. On the way: 5 seconds $1.17 instead of $2.72, 10 seconds $2.32 instead of $5.44, 15 seconds $3.48 instead of $8.16, 20 seconds $4.63 instead of $10.88, 25 seconds $5.79 instead of $13.60.
    • The same clips. 720p with sound, in six shapes: 16:9, 9:16, 1:1, 4:3, 3:4 and 21:9, at the lengths above.
    • Start from a picture. All three take a start picture. On Seedance 2.5 (direct) the clip keeps the picture's own shape, whichever shape is picked, and is charged for the shape it comes out in, never more than the price shown.

    Improvements

    • The price you see is the price you pay. The price the picker shows is what ByteDance bills for the clip, and what you are charged. A clip that fails or is refused costs nothing.
    • If ByteDance is busy, out of credit or unreachable, the clip runs on the version you had before, at its price, and the receipt names the one that ran. If your balance does not cover that price, nothing runs and nothing is charged. A refusal never moves.
    • Busy hours. ByteDance renders a few clips at a time for askr. One that has to wait its turn keeps its credit held while it waits and is charged only when it lands.

    Bug fixes

    Nothing this time.

    Also

    • Real faces. ByteDance refuses a start picture that shows a real person's face, and a prompt or clip that breaks its content rules (W816 on the error codes page). You are not charged, and the other versions are still there.
    • Privacy. The privacy notice and Help's privacy page name BytePlus, ByteDance's own API, behind the Seedance versions marked (direct): it receives the prompt and any start picture with a fixed code made from your account id, serves the clip for 24 hours and keeps the record of the job for 7 days.

    Moved on: Router rebuild

    Announcement to follow.
27. 1.72Answer length30 September 2026Choose how much askr writes: Shortest, Short, Normal, Long or Longest.

    Set the length every new chat starts with in Settings, give one chat its own in Choose a model, or ask for an answer Shorter or Longer from its More menu. It changes how much is written, never how carefully the model thinks or which model askr Pro picks, and Normal is askr exactly as it was.

    New features

    • Answer length. Five lengths on one slider. Shortest is the answer and nothing else, in one or two sentences. Short is the answer first, then a few short bullets. Normal is the usual length for each question. Long is more depth: reasons, examples, steps and the caveats that matter. Longest is everything the question can use, with stories, guides and plans in full.
    • Your default, in Settings. A new Answers group in Settings, Account, right after Appearance, sets the length every new chat starts with, and every chat you have not set on its own. It is kept in this browser; a browser that has never set one starts on Normal.
    • One chat's own length. In Choose a model, right under the askr Pro row and above the models, the same slider sets the length for this chat alone, under askr Pro or with a model you picked. It names your default under it and, when the two differ, offers Make Long my default. A first message from the home page takes its length into the chat it opens.
    • Shorter and Longer. An answer's More menu has Shorter and Longer after Try another model, each naming the length it goes to. Either answers the same message again one step shorter or longer and keeps the chat at that length for the messages after it. They are where Try again is, on the newest answer once nothing is running; Shorter is not offered at Shortest, nor Longer at Longest.
    • You can see it. An answer written at any length but Normal carries the length's name as a small badge in its header, after the model and why, and Details has an Answer length row. While a chat is not on Normal, the askr Pro label in the box reads Pro, Short, a model you picked shows the length after its name, and the + sheet's Choose a model row reads askr Pro, Balanced, short answers.

    Improvements

    • The same words for every model. The length reaches the model as one instruction at the end of its system prompt, the same for every model and every supplier. It comes after everything else the model is told and says it overrides any earlier wording about length. The (direct) versions of GPT-6.1 Sol, GPT-6 Sol, GPT-6 Luna and GPT-6 Astra, and their pro modes, also get OpenAI's own verbosity setting: low for Shortest and Short, high for Long and Longest.
    • Room for Longest. A Longest answer gets the Code door's room: up to 32,768 tokens, or 65,536 on a (direct) version, and its twenty-minute clock with the hold renewed every minute, where an ordinary answer stops at 8,192 tokens and four minutes. A long story, guide or plan finishes instead of being cut off. Shortest to Long keep today's ceilings, since thinking counts inside the ceiling and a small one would leave answers empty.
    • Your words win. A message that asks for another length, in one line or in full detail, gets what it asks for at any setting, and the next message goes back to the chat's length. Code, formulas, tables and anything you ask to have written out come whole, with the length applying to the words around them. Safety information is never left out, and the answer never mentions the setting.
    • Code at every length. On code, Shortest and Short put the code first and whole with a line or two around it, and Long and Longest go on to how it works, how to run it and what to change.
    • Thinking untouched. The length changes how much is written and nothing else: how hard the model thinks, which model askr Pro picks, and whether it searches the web are all as they were. Cheap, Balanced and Performance still decide how strong a model askr Pro picks.
    • Levels about models. Cheap reads A cheaper model and Performance The strongest models and more searching, where they promised short and thorough answers, since how much is written is now Answer length's to say. In a skill at any length but Normal, Cheap no longer asks for a short reply and Performance keeps its checks without asking to be thorough; at Normal a skill reads exactly as before.
    • A truer estimate. With a model you picked, the estimate before you send follows the length. On a new chat's first message a typical answer is counted at 60 tokens at Shortest, 200 at Short, 700 at Normal, 1,500 at Long and 4,000 at Longest (on code, 3,000, and 6,000 at Longest), and a Longest estimate allows for the bigger ceiling. In a chat already under way, the typical answer grows and shrinks in the same proportion.
    • Everyday, not quick. askr Pro's reason for an everyday answer written Long or Longest reads for everyday questions, not for a quick answer.

    Bug fixes

    • A chat on a (direct) model, Claude, GPT or Grok, with web search on or with askr Pro making a picture, clip or sound beside the words, stopped with error W411 before it answered. It now searches and makes things there as it does on every other version.

    Also

    • Normal is unchanged. At Normal a model gets, byte for byte, the request it got before: no instruction, no verbosity setting, the same ceilings and clock. Normal is where everyone starts.
    • Credits. There are no new prices. Output is billed as it is written, so a shorter answer costs less, but the conversation sent with each message still counts, so the saving is smaller in a long chat. A Longest answer holds more credit while it writes, sized on its bigger ceiling, and the unused part comes straight back when it finishes.
    • Shared chats. Each person's length goes with their own messages, so in a shared chat every @askr is answered at the length its writer set, and Try again keeps the length of the message it answers.
    • Stories in full. Write and edit follows the length too. A new piece it writes for you, such as a story askr Pro hands it, is written at the chat's length, so at Longest the story comes in full. Text you give it to edit or fix keeps its length unless you ask for another.
    • Where it is not. Pictures, clips, songs and voices, askr Community, askr Pro's own decisions and askr Mind's notes are as they were. Summarise and Translate keep their own length, since a summary has its own length to choose and a translation follows what you gave it.
    • Help. Answer length has a page: the five lengths, where to set them, Shorter and Longer, what always stays whole, and what it does to the price. The askr Pro levels in Help talk about models only.

    Announcement to follow.
28. 1.71Drawn in code30 September 2026Claude, GPT and Grok can now draw.

    Pick one of their strong models under Image or Video in Choose a model and it writes your logo, icon, diagram, chart or title card as code, which askr turns into a picture or records as a clip, and askr Pro gives them the asks they are good at on every level.

    New features

    • Drawn in code. Eight models draw a picture or animate a clip in code: Claude Sonnet 5.5, Opus 5.5 and Fable 5.1, GPT-6.1 Sol, GPT-6 Sol and GPT-6 Astra, and Grok Build 0.1 and Grok 4.7. Choose a model lists them at the end of Image, under Drawn in code, and of Video, under Animated in code, each with the most one can cost: from $0.06 a picture on Grok Build 0.1 to $1.36 on GPT-6 Astra, and from $0.07 to $1.71 a clip.
    • Sharp words and lines. A drawing is written, not painted, so its words come out as real text, far more reliably spelled than in a painted picture, and its lines stay sharp: logos, icons, diagrams, charts and cards with words on them, and for a clip, titles, moving words, charts that build and logos coming in. It lands as a normal picture or clip, in the chat and in your Library.
    • askr Pro draws too. Ask for a logo, a diagram, a chart or a motion graphic and askr Pro can give it to a model that draws in code: Grok Build 0.1 at Cheap, Claude Opus 5.5 at Balanced and Claude Fable 5.1 at Performance, each with another maker's model behind it if a drawing fails. A photo, a painting, a product shot or filmed footage stays with the picture and video models.

    Improvements

    • Pay for what it uses. A drawing is charged for what the model wrote, never more than the price shown, and one that does not come costs nothing.
    • The chat keeps the model. Pick a model that draws and the chat's next pictures or clips are drawn by it too, a clip at the same shape and length. Try again draws the line again with the same model and, for a clip, the same shape and length, and Try another model lists the models that draw beside the picture models.

    Bug fixes

    • In a chat, Choose a model's Image and Video tabs listed nothing. They now list the picture models and the models that draw, and the clip models for the chat's creator, so you can switch without starting a new chat.

    Also

    • Help. Pictures, clips and songs has a section on pictures and clips drawn in code: which models, what they are good at and what they cost.

    Announcement to follow.
29. 1.7Launch-ready askr30 September 2026askr is launch-ready: the app we built in the open at app.heyaskr.ai is now heyaskr.ai itself.

    One design across the site, the chat and Help; askr Pro picks the model so there is nothing to choose; any chat becomes a shared one with one tap; skills make pictures, clips, songs and whole sites; a Media studio puts it all together; askr Mind remembers you from one chat to the next; and every price is in dollars.

    New features

    • One design. The site, the chat and Help wear one header, and the logo always goes home. On a desk the side column stays open, on a phone the askr logo is the menu, titles sit on one line, and the moon under the box is a planet with a body. Light is the default, with Dark and System in Settings.
    • Just ask. The home page plays ten examples made for real on askr, on a loop: a site, a picture, Look things up, a song, askr Collab, a clip, Money, askr Mind, Translate and Code. The prompt types itself, askr Pro shows which models it picked and what each one did, and the real result lands. Under them is the Ask anything box on the moon, then a fuller How it works and the common questions. The first send opens a chat on askr Pro, Balanced.
    • Every chat can be shared. A chat is an askr Collab waiting for people: Invite people to join brings them in, and the chat's shared credits read as one number. An older chat opens as it was and carries on, shareable, from its first new message.
    • askr Pro, no pickers. Under askr Pro the box has no pickers. The + opens three rows, Add a photo or file, Choose a model and Load a skill, the same on the home page and in a chat, and what you choose sits over your message as one chip you can take away.
    • Making skills. Load a skill opens with Make a picture, Make a clip and Make a song, each with its own small setup (how many, the shape, the length, sung or instrumental), then Build a site and Ask about a file, then the ten askr Skills.
    • Media studio. Also in Load a skill: lay your pictures out as an album or a collage, or put clips, pictures and a song together into one video: drag them into order, trim a clip, set how long each picture shows and where the song starts, then watch it back. It is made in your browser, saved to your Library at no charge, or downloaded.
    • Sites you can see. Every answer with a site has Preview, whichever way askr Pro answered it, and each version previews the site as it stood after that answer, so a follow-up shows the change in it. A site that needs live data runs in the Preview on sample data, an answer that was cut off offers Continue, and a project only a server can run offers Make a version that runs here.
    • Chats and Library. One Chats list, every chat newest first, with a search that reads inside every chat's lines, and one Library of everything made in your chats, with your uploads beside it.
    • Memory across your chats. With Remember me across chats on, in Settings, Memory, askr Mind carries what it learns about you from one private chat to the next. It learns only from chats of your own, and a shared chat reads it only when you turn on Share my memory here in that chat, never a public one.
    • Settings in four. Account, Credits, $ASKR and Memory. Download or delete your chats under Account, and Send credits sits under Credits.
    • Dollars, not credits. Your balance, every answer's price, Add credits, Activity, Pricing and Help all show dollars. A dollar paid is still a dollar of credit.
    • One Add credits sheet. Crypto first, SOL, ETH and USDC with every other coin behind Other coin, then card, and it brings you back to where you were. A send that needs more credit becomes a card in the chat with Run it now for when the credit lands.
    • Sign in in one step. One field, Email or @username, with a wallet as a link beside it, and it can open over the chat.

    Improvements

    • Pricing on one page. What an answer, a picture, a clip and a song cost at each askr Pro setting, with what an $ASKR holder pays under every price; the 95% holder discount as it works today, and the discount on askr's fee that Token Flywheel Proposal #1 proposes beside it, marked Proposed; and every model, filtered by General, Picture, Video and Audio and then by maker. /models, /calculator and /methodology land there.
    • The token page. The live price chart, a pie of who holds the supply, the locked team tokens and their schedule, the 95% holder discount, and the fee discount the proposal would bring.
    • Help in depth. Every article is rewritten in plain words: the short version first, then what to tap. Every error reads in plain words with the one button that fixes it.
    • Fewer pages, no dead links. How it works is part of the home page, About and Contact are Talk to us, and the guides and comparison pages lead home. app.heyaskr.ai sends every link to the same page here.
    • The feedback bar waits. In the chat it stays hidden until your first answer, and on the site it is one quiet line.
    • Newer models. GPT-6.1 Sol on GPT (direct), with the pro mode of GPT-6.1 Sol, GPT-6 Sol and GPT-6 Luna, and Claude Sonnet 5.5 on Claude (direct). askr Pro runs GPT-6.1 Sol and Gemini 3.8 Flash in their older siblings' places at the same prices, and Claude Sonnet 5.5 is a chat's own model unless you pick another.
    • More clips and voices by fal. Seven more clip models at fal's prices, each listed under the model it stands for and from 13 to 42 percent under it, and the ElevenLabs v4 and v4 Turbo voices.

    Bug fixes

    • A chat's own model no longer falls to the cheapest one, and askr Pro's Balanced answers run on Claude again, after the gateway renamed its models on 28 September.
    • A plain question under askr Pro gets the answer, not a lesson: the Learning skill starts only when you ask to be taught, tested or walked through something, and every skill keeps to a length you ask for.
    • A Look things up answer that searched the web draws its sources and follow-ups again instead of raw code.
    • A chat's model no longer talks about how askr routes a message, or says you picked a route.
    • A Code answer can run to 32,768 tokens, 65,536 on a (direct) version, where it was cut off at 8,192.
    • An answer whose model said it had finished is no longer marked cut off because an error came through as well.
    • A song asked for in seconds is made at that length where its model can make one.
    • Make a song is offered when you are signed out, like a picture or a clip, and sending asks you to sign in, where it said songs could not be made.

    Also

    • What went. Compare, temporary chats and projects are gone from the app. Comparisons you made still show, chats already marked temporary still work, and nothing anyone made was deleted; a chat with people in it is the way to work together.
    • Developers. The developer pages and the playground are gone from the site. API keys you made are listed under Settings, Account, with Revoke.
    • Token Flywheel Proposal #1. The proposal sets out what $ASKR is for and how the flywheel would run, step by step, with a thumbs up, a thumbs down and a comment box. Nothing in it is live yet.
    • Terms. The terms name heyaskr.ai alone for the deposit address and say a key is revoked in Settings, under Account.

    Finished: askr ProMoved on: askr SkillMoved on: askr Mind

    Announcement to follow.
30. 1.61Direct xAI routing28 September 2026Grok now runs on xAI's own API too.

    Pick the version marked (direct) under Grok 4.7, Grok 4.6 or five more, and it costs xAI's list price, shows it is thinking from the first second, has twice the room to think, and keeps the conversation cached between turns so a follow-up costs a quarter to read. The gateway versions stay where they were.

    New features

    • Grok (direct). Grok 4.7, Grok 4.6, Grok 4.5, Grok 4.3, Grok 4.20, Grok 4.20 Multi-Agent and Grok Build 0.1 straight from xAI, listed under each model as its (direct) version, in chats and in rooms.
    • Room to think. A direct reply can run to 64,000 tokens, thinking included, and the chat shows the model thinking from the first second instead of a blank wait. Grok 4.7 and 4.6 take the four thinking depths; 4.5 the first three.
    • Cached conversations. The transcript is cached between turns automatically, so a follow-up reads the history at a quarter of the input price; there is no charge for writing the cache.
    • List price, to the tick. A direct version costs what xAI charges, and the receipt is xAI's own figure for the call, cache discount included, not an estimate from a rate card. On six of the seven that is 5.5% under the gateway's copy; on Grok 4.7 the gateway's copy is the cheaper one, and the picker lists it first.

    Improvements

    • Receipts name the source. A run on a direct version says (direct) on its receipt, and the ledger records where it ran.
    • Pictures as pictures. A direct version reads a picture you attach as the picture. PDFs and other files go as their text, since xAI does not take a file.
    • If xAI is busy, out of credit or unreachable, the turn runs on the gateway's version at its price, and the receipt names the one that ran. A refusal never does.

    Bug fixes

    Nothing this time.

    Also

    • Content rules. The refusal pause covers this lane on its own, as it does the other two: an account xAI refuses five times in a day loses the Grok (direct) versions for a day (W416, on the error codes page), and every other model keeps running.
    • Privacy. The privacy notice and the docs privacy page name xAI behind the Grok versions marked (direct): it receives the request with a fixed code made from your account id, is asked not to store the conversation, keeps requests up to 30 days to check for abuse and does not train on them.

    Moved on: Router rebuild

    Announcement to follow.
31. 1.6askr Pro27 September 2026A model you never have to pick.

    A new chat opens on askr Pro, Balanced: it reads each message with the room's memory and the tier, picks the door and the model, starts a skill when a line calls for one, makes the pictures, clips and sounds the line asks for beside the words, packages the lot around what you asked with a preview beside the room, and shows every pick under your line with Change.

    New features

    • askr Pro in the picker. The first row of the model picker in a chat or a room, with Cheap, Balanced and Performance beside it. Choose it and askr Pro picks the door and the model for every line at that tier, and the picker's trigger shows the mark and the tier word. The tier is a ceiling and a floor: Cheap never runs a dear model, Performance always picks the best there is, and a hard question moves up to the tier's reasoning models, never out of the tier, whatever the message says.
    • On from the start. A new chat opens on askr Pro, Balanced, and so does the Ask composer on the home page, whose first send makes a room on it. Pick a model or a skill by hand and that sticks; the tier, or the model you picked instead, is remembered on that device. In a room the creator's pick sets askr Pro for the room and anyone else's holds for their own lines.
    • What it reads. The message and what is attached to it, the last lines of the thread, the Building and Made lines of the room's memory, the tier, and the doors the room has on. One call to a decision model that answers with probabilities, not words, says how likely each door is, which of thirteen kinds of task the line is, whether it wants the web, how hard it is, whether it carries on the last one and whether it also wants something made; plain code turns the numbers into a door and a model, so nothing you type can talk it out of the tier. The decision has two and a half seconds and runs before the credits are held; when it is slow or down the line still runs, on the room's model through Ask, and the pick says so.
    • Every pick shown. Under every line askr Pro decided, one line says what ran and why: askr Pro: Claude Sonnet 5 for writing, Balanced; askr Pro kept Claude Sonnet 5, this continues the last answer; askr Pro: Look things up skill on Claude Sonnet 5, Balanced. Change beside it opens the picker; pick a model by hand and the line runs again on it, askr Pro is off for the room from then on, and the model you picked is the room's model. While it works the room says askr Pro is choosing, then which model it picked, and a clip or a sound names the renderer or the voice on its wait row.
    • No doors to choose. With askr Pro on the door chips under the message box are gone: it picks Ask, Code, Image, Video or Audio for every line itself, and there is no estimate beside Send, since the price is not known before the pick. A model picked by hand brings the chips and the estimate back. A line that opens by naming a picture, a photo of, a portrait of, a logo for, always paints, and a read-aloud ask in any of the common languages is read aloud.
    • Made beside the words. When a line wants a picture, a clip or a sound as well as words or code, a site and its branding, a plan and its poster, the answering model makes them itself as it writes: it writes each prompt, each one runs through the room's own door on the tier's painter, renderer or voice, every made thing lands under the answer as its own line, and making a picture shows while it does. Nothing sends you to a door to do it yourself.
    • Packaged around the ask. Code carries the pictures it was made with, by their addresses, instead of a file name you would have to supply. After a run that made things askr Pro checks once, off the turn, whether everything asked for is there and connected; whatever is missing is made in one more run, and a closing card under the last reply sums the package up, Your dog grooming site: 2 pages, logo, hero, gallery, 2 promo clips, 3 blog posts, with the bill for the whole line, split by clips, pictures, sounds and code or words, a clip still rendering counted when it lands.
    • The preview takes the room. When a thread has a page to show, on a desktop the rail folds away and the panel takes the right half at full height, Preview first, with a small button that reaches Room settings and People; on a phone one switch moves between the chat and the full-screen preview, with Chat to come back. The page is assembled from the answer's files, its stylesheets and scripts inlined by name, even when the fences carry no names, and every picture, clip and sound made in the thread is put into the page before it is shown. Anchors scroll, a link to another page of the answer shows that page with Back, a link to a made picture, clip or sound opens it in the room, an outside link opens a new tab, and the frame never navigates away.
    • Files, and Made here. A Files tab beside Preview and Code lists everything the thread made, pictures, clips, sounds and code files, each with Open and Save, a rendering clip saying so. A Made here card in the rail, and at the top of the Room sheet on a phone, shows the newest six pictures and clips, solo or collab, with All opening the Library.
    • Code is not a wall. In every answer, in rooms, chats and skills, the fences fold into one row, Your code, 3 files, closed until you open it; the files are listed by name with Copy and Save and open one at a time. A short snippet with no file name stays inline.
    • Skills, without asking. A line that calls for one of the ten skills, looking something up, writing, learning, fixing something, health, translating, money, ideas, a summary or shopping, runs with that skill's instructions and screen in the same thread: the sources, steps, cards or calculators arrive under the answer, and the pick says which skill ran. A thread that already carries a skill keeps its own.
    • A follow-up keeps the model. When a line carries on the last task, the model that answered last answers again, as long as it is still live and inside the tier; the pick says kept. A new task picks afresh.
    • Asked before a dear door. A picture or a sound needs a confident read, a clip a more confident one, and a follow-up in a thread whose last line came through that door a lighter one. Under the bar the line goes through Ask with a chip under it, Make it a picture instead, or a clip, or a sound, that runs the same line through that door. Ask and Code need no bar.
    • Never stopped in your tracks. When a supplier declines a picture, a clip or a sound under its content rules, askr Pro rewords the request once, the brand, company, real person, franchise or copyrighted work replaced by the kind of thing it is, says so in the room, and tries again. A second refusal becomes an offer on your line, what could not be done and what would be done instead, with a chip, Yes, make the sound that way, that runs it. Inside a written answer the model rewords and tries again itself. With askr Pro off a refusal reads as before.

    Improvements

    • Search the web, when a line wants it. When the decision says a line wants the web, Search the web is on for that line, under the room's usual rule.
    • A painter per line in rooms. The Image door in a room takes a painter picked for that one line, from the composer or from askr Pro, without changing the room's model.
    • A file on the line is read. A line with a document or a data file attached that reads as an unsure Code pick runs through Ask instead, where the model reads the file.
    • The tier words. The first tier reads Cheap, in the picker, the skill tier pill and the docs.
    • The picker on a phone. The trigger shows the mark and the tier word alone, so it fits the narrow pill.
    • Admin. An askr Pro page in the admin area: decisions in the last day and week, fallbacks, kept, changed by hand, media doors picked, average and p95 time, cost, by door, tier and class, the confidence buckets, rooms on askr Pro by tier, and the newest decisions.

    Bug fixes

    • In a room, picking a painter for one line answered W201; the line now runs on the painter it named.
    • A picture call has four times the output room it had, and a picture cut at the budget (W513, four times in a row in one room) is asked for once more on the same hold, so the check after the run no longer has to make them all again.

    Also

    • Docs. askr Pro has a page: what it reads before each pick, how it decides, the rules, what it makes beside the words and how it packages it, what happens when a supplier refuses, what it costs and where it is not offered; askr Skills says a skill can start on its own and askr Collab names the picker's first row.
    • What it costs. Each decision costs a tiny fraction of a cent, metered under its own source and charged at zero while askr is feeless, and so is the rewording of a refused request. The model that answers, and every picture, clip and sound it makes, is priced as always; nothing asks before a dear pick, the credits show under the answer as on every answer, and a line that made things shows the whole bill on its closing card. A line that asks for a site, a clip, a song and four pictures at Performance can run to several dollars, and the card says what it came to.
    • Where it is not. A skill thread you opened yourself keeps its tier pill, since the tier there already picks the model; Compare names its own models and never takes askr Pro.
    • Privacy. The decision reads your line the way any model you run does, and the record askr keeps of each decision, for the admin page and for version two, carries no message text.
    • Measured before it shipped. Over 260 test lines, in six languages, with files, follow-ups and lines that try to name a dearer model, the door was right 98 to 99 times in 100 at every tier, the web read 97 to 98, and a decision took about 280 ms in the middle and under 400 at the 95th, for a fraction of a cent. A room driven through every door, the tier switch, the chips, Change, askr Pro off and the refusals passed 49 checks of 49.
    • What is next. Version two builds the ladder around you, from what you reach for, ask and keep: the askr Pro step on the roadmap.

    Moved on: askr Pro

    Announcement to follow.
32. 1.51Direct OpenAI routing27 September 2026GPT now runs on OpenAI's own API too.

    Pick the version marked (direct) under GPT-6 Astra Pro, GPT-6 Sol or five more, and it costs OpenAI's list price, shows it is thinking from the first second, has twice the room to think, and keeps the conversation cached between turns so a follow-up costs a fraction to read. The gateway versions stay where they were.

    New features

    • GPT (direct). GPT-6 Astra Pro, GPT-6 Astra, GPT-6 Sol, GPT-6 Luna, GPT-5.3-Codex, GPT-5.4 Mini and GPT-5.4 Nano straight from OpenAI, listed under each model as its (direct) version, in chats and in rooms.
    • Room to think. A direct reply can run to 64,000 tokens, thinking included, and the chat shows the model thinking from the first second instead of a blank wait.
    • Cached conversations. The transcript is cached between turns, so a follow-up within half an hour reads the history at a tenth of the input price. Writing the cache on the first turn costs a quarter more on GPT-6; on a long chat the reads take most of the input cost off.
    • List price. A direct version costs what OpenAI charges, 5.5% under the gateway's copy on every token.

    Improvements

    • Receipts name the source. A run on a direct version says (direct) on its receipt, and the ledger records where it ran.
    • PDFs as files. A direct version reads a PDF as the file itself, its text and its pages, and pictures as pictures.
    • If OpenAI is busy, out of credit or unreachable, the turn runs on the gateway's version at its price, and the receipt names the one that ran. A refusal never does.

    Bug fixes

    Nothing this time.

    Also

    • Content rules. The refusal pause now covers each maker's direct versions on their own: an account OpenAI refuses five times in a day loses the GPT (direct) versions for a day (W416, on the error codes page), and Claude (direct) keeps running, and the other way round. Every other model keeps running.
    • Privacy. The privacy notice and the docs privacy page name OpenAI behind the GPT versions marked (direct): it receives the request with a fixed code made from your account id, is asked not to store the conversation, keeps requests up to 30 days to check for abuse and does not train on them.

    Moved on: Router rebuild

    Announcement to follow.
33. 1.5askr Mind26 September 2026What you told one model is there for the next.

    Every chat and every room keeps a short memory that goes first into every model's prompt, with one more about you across chats, and all of it is yours to read, pin, edit, forget or switch off. A new chat is now a room of one, the skills and Compare work inside rooms, and a project can hold a team, so a room knows what the other rooms in it are doing.

    New features

    • A memory every model shares. After every answer askr Mind notes, in one line each, who is here, what is being built, what was decided, what is open, what was made and what each person prefers. The block goes first into every prompt, every model in a room reads the same one, and the chat says Using askr Mind while a model is being briefed. Nothing is rewritten and nothing is deleted on its own: entries are only added or replaced.
    • Yours to keep or not. The askr Mind card in Room settings lists what the room remembers, click to edit, Forget on any line, Forget everything with a second press, and Remember this room for the creator. Pin beside Reply puts a line into the memory. After an answer a Mind noted chip says what was just remembered.
    • Made, with captions, and Prompt used. Every picture, clip and sound made in a chat or a room is listed under Made; a picture gets a one-line caption of what is in it. When you ask a door for something, the prompt the generator gets is written from the memory and your request, and Prompt used under the result opens to the exact words, with the picture it built on. Ask for a made picture, moving, and it goes to the clip as the start frame.
    • Recall. When a thread holds more than the prompt carries, the words of your request are searched across every line of the room and up to five earlier lines that match are handed to the model, each with who said it.
    • About you, across chats. One list of who you are, what you build and how you like things, learned from your chats and put in front of every one of them. Settings, Memory shows it, with Remember me across chats, the entries, and Forget everything. A temporary chat, from the dashed bubble in the message box, keeps askr Mind out of that chat entirely.
    • Projects. One memory shared by the rooms and chats you put in it: make one under Settings, Memory, pick it beside a chat's title or in Room settings, and every prompt in those rooms and chats carries it. Shared brings what askr knows about you into its rooms too; Walled off keeps you out.
    • Your memory in a room, your choice. Bring my memory into this room, in Room settings, is each member's own switch for that room. A public room never takes it.
    • Team projects. A project has people in it: invite by @ under Members on the Memory page, up to twelve, and anyone in it can put their private rooms in. The project's memory then says in one line per room what each room is doing, keeps every picture, clip and sound made in any of them with the room it came from, and a question in one room reaches the lines of the others: ask what is happening in the front-end room and the model names the room and the work. A picture asked for in one room can start from a picture made in another. A public room never joins, and leaving takes your rooms out with you.
    • Recall by meaning. Every room line and memory entry is embedded, so a question about the thing we said about pricing finds it without the word pricing in it, beside the search by word. A long room stays reachable to its first line.
    • Made from memory, with references. The prompt a picture, clip or sound door receives is written from what the room remembers, and the picture it builds on is handed to the run: a clip starts from the picture, an edit starts from the picture. Prompt used under the result shows the words, and a Prompt from what the room remembers preview appears under the message box while you type in a media door.

    Improvements

    • A room of one. A new chat is a room with one person in it, so a chat can do what a room does: threads, New thread from any line to continue from there, Re-run on the newest answer, a model per line from the composer, receipts that say the time taken and the pot that paid, and Export. It pays from your balance as a chat always did, free credits first, then given, then bought with the holder discount, and a run started alone settles alone even if you invite someone while it runs. Recent lists your rooms.
    • Skills and Compare in rooms. A skill opens as a room whose thread carries the skill: the banner names it, Change starts a new thread under another skill, and the settings live on the thread so everyone in the room sees the same ones. Compare beside the model picker sends one line to two to four models side by side; every run holds its credits before any starts, Pick keeps one, and later lines and the memory read the picked answer only.
    • Rooms say who spoke and what was made. A model in a room reads each earlier answer with the model that wrote it, and a picture, clip or sound as a line that says which door made it, with its caption. It is told which doors the room has and to say which one makes what you asked for, and never to pass a picture off as its own.
    • Estimates that match receipts. The price beside Send in a room adds what the memory and a skill's brief put in front of the model, so the number before matches the receipt after.
    • A room of one reads as a chat. No purse controls until someone is invited, tools and receipts in sentence case, a one-time note under the first answer saying the room remembers, and the project the room is in named on its title line.
    • Continue in a room. An older chat gets Continue in a room, on the chat and in History: its lines and pictures become a room of one with the memory built at once.
    • Every door starts a room. The Images, Video and Audio doors on the home page open a room set for that door on the first send, the start frame riding along.
    • A memory that stays tidy. A near-duplicate of a line already remembered is refused, Who is here says what each person does, a Pick sets the thread's model, clips and sounds get a caption too, and a room folds at most once every twenty seconds however busy it is.
    • Put this in a project? A new room whose plan resembles one of your projects offers to join it, once.
    • Admin. An askr Mind page in the admin area: entries by scope and heading, the writer's calls and their cost, the rooms with the most memory.
    • Handed on in rooms too. A Claude model that thinks through its whole budget in a room and never writes is handed to GPT-6 Sol under its own name, as the chat has done since 1.395, and nothing is charged for the first run.

    Bug fixes

    • A room of one short of credits for a picture or a clip now says so as a chat does, with the balance and Add credits, instead of the room wording.
    • A sound made in a chat is described to the model as a sound; it used to be described as a picture.
    • A picture asked for in a young room with the words make them, right after a message about three pictures, reached the image model as those bare words; the prompt writer now reads the last lines even before the room's first memory entry.
    • An image model that answers a room in words instead of drawing is asked once more, plainly for a picture, on the same hold.
    • Re-run on an answer ran the wrong line: it now runs the question the answer sits under.
    • A wrapped caption was cut at its line break, and a writer that ran out of room kept a half-written last line; both now come through whole or not at all.

    Also

    • What it costs. askr Mind's own calls are metered under their own source and charged at zero credits while askr is feeless. They never hold a turn up, and a failure in askr Mind never fails a turn.
    • Docs. askr Mind, askr Collab and askr Skills each have a page.
    • Privacy. The privacy notice says what askr Mind keeps, that a temporary chat is neither read from nor written to, that your memory enters a room only when you switch it on there and never a public one, and that deleting a chat, a room or your account deletes its memory with it.

    Finished: askr Mind

    Announcement to follow.
34. 1.48ASKR Connect improvements27 September 2026Verify holdings is smoother for apps that buy through a router, as pump.fun does.

    When a buy like that cannot show whose wallet it is, the sixty seconds to sell start at once, and the sale you make is the proof.

    New features

    Nothing this time.

    Improvements

    • Straight to the sale. When a buy is turned down, the page shows the sixty seconds to sell right away, with Done, I sold. There is no button to press first, and a sale made after the verdict counts.
    • The sale step is part of your check. Checks are limited to three an hour, and the sale step used to take one of them. It now belongs to the check you started, so it never uses one up.
    • A contract says so at once. A sale from an address that is a contract cannot prove it either, so the page says that straight away and points you to Connect wallet, where signing does.

    Bug fixes

    Nothing this time.

    Announcement to follow.
35. 1.47Model audit26 September 2026A user tested three of our models and said they were older than their labels.

    We tested all 383, published everything, and shipped what the test taught us.

    New features

    • The model audit report. Every model on askr against its label: the user's test, all 383 chat models with their verdicts, what we changed, and how to check for yourself.

    Improvements

    • Every model gets today's date. One line in front of every chat and room turn, so a model no longer calls last year's news "not yet happened".
    • Cannot tell yet on the catalog. The 55 models the audit could not settle say so on their card, with the reason.
    • Old names off the list. gpt-3.5-turbo-0613, gpt-3.5-turbo-16k and wizardlm-2-8x22b answer like newer models, so they are off until our supplier says what serves them. Eleven models that no longer answer are off too, and a delisted one is refused instead of run.

    Bug fixes

    Nothing this time.

    Also

    • Refund. Everyone who ran Grok 4.7 got back what its route's hidden prompt tokens cost them, with no need to ask.

    Announcement to follow.
36. 1.46Rounding off26 September 2026A tidy-up release.

    A public collab you share now brings you referrals, only one sound plays at a time, every error code explains itself when you tap it, and GIFs move in collab chats.

    New features

    • Public collabs refer. Share a public collab you made, and anyone who makes their askr account from it becomes your referral: 10% of the credits they add comes to you, the same as your own link. An account that already exists, a private room, or a sign-up that came with someone's code does not count. Referrals has the rules.
    • GIFs in collab chats. A GIF sent to a room moves, at its own size, so a live chat can have some fun.

    Improvements

    • Every error code explains itself. The code at the end of an error is now a chip: hover it, or tap it on a phone, and it says why it happened and what to do, with a link to its row on the error codes page.
    • One sound at a time. Play a song, a voice or a clip and whatever else was playing pauses, in chats, in collabs and in your library. Read aloud does the same.

    Bug fixes

    • Adding credits to a room, making one or joining one now shows the code when it fails, like every other error.
    • A collab run or reading short of credits now carries a code (W301, W825).
    • A reading that is too long (W821) and a start frame that is too large (W605) showed their code twice; they show it once.
    • Read aloud on a second reply while the first was reading stopped both; the second now takes over.

    Also

    • Docs. Every code on the error codes page has its own link, and W914 (the card checkout box) is listed. Referrals covers public collabs.

    Announcement to follow.
37. 1.45Voice, music, sound25 September 2026Audio is the fifth door.

    Have a voice read your words, ask for a song, or describe a sound and get it, in chats and in collabs, priced per character, per song or per second like everything else.

    New features

    • Voice. Twelve voices read what you type: ElevenLabs Turbo 2.5 by fal at $0.05 per 1,000 characters, ElevenLabs v3 by fal at $0.10, and ElevenLabs Flash 2.5 and v3 through the gateway. Up to 5,000 characters a reading (3,000 on v3).
    • Music. Describe a song and get one with lyrics: Lyria 3.5 by fal at $0.10 a song, Lyria 3 Pro at $0.08, and ElevenLabs Music by fal at $0.60 a minute for a piece between 30 seconds and five minutes.
    • Sound. A door creaking, rain on a tent, a bell: ElevenLabs Sound Effects by fal at $0.02 for ten seconds, one to 22 seconds.
    • In collabs too. An Audio pill in the room composer; @askr with the words or the description and the room hears it, paid from the room balance, kept in the room library.

    Improvements

    • A player on the line, lyrics under a song, a Made with AI line under every piece, Download on the receipt, and the library keeps your audio beside your pictures and clips.
    • Priced before you send. The button says what a reading, a song or a sound will cost, and a refused request costs nothing.

    Bug fixes

    Nothing this time.

    Also

    • Docs. Make audio covers the models, voices, lengths and prices; the codes W820 to W829 are on the error codes page.
    • Privacy. The privacy notice says what fal receives for a reading, a song or a sound: the text or the description and the name of the voice, kept under fal's terms, used to train nothing.

    Announcement to follow.
38. 1.44Public collabs25 September 2026A collab can now be public: share the link and anyone can watch it live, no account needed, and anyone with an account can join, up to a hundred people on one balance.

    Private stays the default, and the creator now decides what a room can do.

    New features

    • Public collabs. On the People card, or when you start a room, pick Anyone with the link. The room, its answers as they stream, its pictures and clips and its library are then open to anyone who has the link, signed in or not, with a Join button for those who are. Invite only puts it back the way it was; whoever joined stays.
    • Join from the link. A public collab takes anyone with an account and an @, up to 100 people (a private one keeps 12). Past that the link is view-only. Sign in to join brings you straight back to the room.
    • What a room can do. Room settings has a Can use row: the creator turns Ask, Code, Image and Video on or off for the room, and the composer offers only what is on.
    • Room settings for everyone. Every member, and every viewer of a public room, sees the settings and the room balance; only the creator changes them or adds credits.

    Improvements

    • A shared link looks like the room. A collab link unfurls with the askr mark and the room's title on the card; a private room's card names no title. Every other page of heyaskr shares with the brand card too.
    • The room's side. Room settings, then Room (the threads, renamed), then People, from the top of the screen and building downwards as settings open, threads are made and people come in.
    • Copy link beside the choice, and the New room form asks who can see it.

    Bug fixes

    • Public, and how many are watching, on the room's title line; the phone's sheet offers Sign in to a reader with no account.

    Also

    • Privacy. The privacy notice says what public means for a collab: anyone with the link can read it, watch it live and play what was made in it, without an account.

    Announcement to follow.
39. 1.43Direct Anthropic routing25 September 2026Claude now runs on Anthropic's own API.

    Pick the version marked (direct) under any Claude model and it thinks as long as it needs, shows its reasoning as it goes, costs Anthropic's list price, and keeps the conversation cached between turns so a follow-up costs a fraction to read. The gateway versions stay where they were.

    New features

    • Claude (direct). Opus 5, Sonnet 5, Fable 5.1, Opus 5.5 and Haiku 4.5 straight from Anthropic, listed under each Claude model as its (direct) version, in chats and in rooms.
    • Room to think. A direct reply can run to 64,000 tokens, thinking included, and the reasoning streams while it happens instead of a silent wait. The handoff to another model when Claude ran out of room no longer applies to these versions.
    • Cached conversations. The transcript is cached between turns, so a follow-up reads the history at a tenth of the input price or less. On a long chat that takes most of the input cost off.
    • List price. A direct version costs what Anthropic charges, 5.5% under the gateway's copy on every token.

    Improvements

    • Receipts name the source. A run on a direct version says (direct) on its receipt, and the ledger records where it ran.
    • Native PDFs. A direct version reads a PDF of up to a hundred pages as the file itself, and pictures as pictures.
    • If Anthropic is busy or unreachable, the turn runs on the gateway's version at its price, and the receipt names the one that ran. A refusal never does.

    Bug fixes

    Nothing this time.

    Also

    • Content rules. An account that a direct version refuses five times in a day loses the direct versions for a day (W416, on the error codes page). Every other model keeps running.
    • Privacy. The privacy notice and the docs privacy page name Anthropic behind the versions marked (direct): it receives the request with a fixed code made from your account id, keeps it for 30 days under its commercial terms and does not train on it.

    Moved on: Router rebuild

    Announcement to follow.
40. 1.42Video by fal25 September 2026Seven versions of the video models you already use, served by fal at its list price: Kling 3.0 at less than half today's price, Veo 3.1 in place of Veo 3, and Wan 3.0 and Seedance 2.0 at about 13% less.

    Pick them in the video picker under the model they belong to, labelled "by fal". Same models, lower price, nothing else changes.

    New features

    • Kling 3.0 at less than half price. Kling 3.0 Pro by fal renders a 5-second 1080p clip for $0.84 instead of $1.93, and Kling 3.0 Standard for $0.63 instead of $1.45. The 10 and 15 second clips are priced the same way.
    • Veo 3.1 is here. Veo 3.1 and Veo 3.1 Fast come in as versions of Veo 3: an 8-second clip with sound for $3.20 instead of $3.68, or $1.20 instead of $1.38 on Fast.
    • Wan 3.0 and Seedance 2.0 by fal. About 13% less than today: a 5-second Wan 3.0 clip for $0.50 instead of $0.58, Seedance 2.0 for $1.52 instead of $1.74 and Seedance 2.0 Fast for $1.21 instead of $1.39.
    • Sound on every clip, and a start frame on all but Veo 3.1 Fast, with the same lengths and shapes each model offers today.

    Improvements

    • The price you see is the price you pay. A fal render is never charged above the quote the picker showed. If fal bills more, askr absorbs it.
    • A queued render keeps its hold. fal queues work when it is busy, so a render keeps its credits held for up to two hours while it waits, instead of the hold lapsing at twenty minutes.
    • If fal cannot take the job, the render runs on today's route at today's price, and the receipt names the model that ran.
    • Clearer names in the picker. A supplier's newer model sits under the model it replaces by its own name, as Veo 3.1 by fal under Veo 3, and a version that only starts at a longer length says what the model charges for the same clip.

    Bug fixes

    Nothing this time.

    Also

    • Privacy. The privacy notice and the docs privacy page name fal behind the versions labelled with its name: it receives the request, with your prompt and any start frame, and keeps it for 30 days and the clip for at least 7. It makes no zero-retention promise, so pick the plain version when that matters.

    Moved on: Router rebuild

    Announcement to follow.
41. 1.41Pay by card24 September 2026Add credits by card, Apple Pay or Google Pay as well as crypto.

    The price you pick is the price you pay, tax included, and the credits land as soon as the payment clears. Card credits are for your own use: to keep fraud and money laundering out, they cannot be sent, cashed out or turned into crypto.

    New features

    • Add credits by card. Add credits opens on Crypto with Card one tap away, and Balance & deposits now has Add credits by crypto and Add credits by card. Pick $5 to $100 or your own amount, tick that you want the credits now, and Checkout opens Stripe's secure page. A dollar paid is a dollar of credit, the same as crypto, and the smallest payment is $5.
    • The ways to pay where you are. Cards, Apple Pay, Google Pay and Link work everywhere, with local options such as Klarna, Pix and Samsung Pay depending on your location. The price shows in your own currency.
    • Tax included. The price you pick is what you pay: any VAT or sales tax is inside it, and askr adds no fee.
    • Credits land at once. Back on askr the balance fills as soon as Stripe confirms the payment, and a confirmation email follows. Card credits get the $ASKR holder discount, like any credits you buy.

    Improvements

    • Card credits are for your own use, on purpose. We are against fraud and money laundering, so credits bought by card follow the rules below. They keep a stolen card, or money that needs hiding, from passing through askr. Credits bought with crypto work as before.
    • They cannot be sent to anyone. Send credits refuses them, so a stolen card cannot be turned into credits in someone else's account. Your balance shows how many of your credits came from card.
    • They cannot be cashed out or turned into crypto. No credit on askr can be withdrawn, so askr is never a route from card money to cash or coins.
    • A collab room is not a way round it. Card credits can fund a room you share, and whatever the room hands back returns to you as card credits, still unsendable.
    • No chargebacks on credit you used. Used credit is not refunded to the card, and a chargeback on it is contested with the account's usage record: the purchase, your agreement at checkout and the requests that spent it. A run that fails is refunded in credits, as always.
    • A dispute pauses the account. While a card payment is disputed, the account cannot send credits or buy more by card. If the dispute is lost, credits from that payment come back off the balance, and anything already spent is owed before the account can run models again.
    • No referral bonus on card purchases. Signing people up with stolen cards earns nothing.
    • askr on Stripe's page. The checkout carries askr's name, logo and colours, and your card statement reads LINK.COM* HEYASKR.AI.

    Bug fixes

    Nothing this time.

    Also

    • Terms and privacy. Card purchases are sold through Link, Stripe's merchant of record. The terms now cover card purchases, refunds and disputes, and the privacy notice names Stripe and Link. Add credits in the docs covers paying by card.

    Announcement to follow.
42. 1.4askr Skill24 September 2026Ten skills behind the Ask tile, each a chat with its own screen that builds itself around what you ask, and starts before the answer lands.

    The model writes plain words and small pieces of structured data; your device draws them as cards, calculators, steps and sources, checks them, and does the maths.

    New features

    • Ask opens a chooser. Tap Ask and pick Ask anything, the plain chat as before, or one of ten skills: Look things up, Write and edit, Learn and homework, Fix it and how-to, Health and fitness, Translate, Money, Brainstorm, Summarise, Shop and compare. Each opens a chat built for that job, on the model that suits it, with its own message box hint and its own settings.
    • A screen per skill. Beside the chat on a desktop, a sheet on a phone, each skill draws what it made. Look things up shows every page it read as a card with the exact words that back each sentence. Write and edit keeps the draft beside the chat, with the changes marked, a word count and reading age worked out on your device, and one-tap rewrites. Learn and homework reveals a worked solution one step at a time with a hint on each, checks your arithmetic on the device, and marks a quiz there too. Fix it shows a Fix card: what you need before you start, the difficulty, the steps easiest first, a did-this-fix-it check under the current step, and any risky step locked until you have read its warning. Health and fitness draws each test result on a bar against the range printed on your report, builds your appointment list to tick off on the day, and shows a care card that says when to get help. Translate gives a two-part card with tappable words and other choices, a clarifying question when the meaning could go two ways, and photo translation for menus and signs. Money runs loan, mortgage, savings, debt and budget calculators on your device with the maths shown, and reads a statement photo into spending by category and the charges that repeat. Brainstorm deals ideas as cards to keep, drop, combine and build on, with a lens to filter them. Summarise gives an overview card with the quote behind each point at the length you choose. Shop and compare asks what matters first, then puts the options side by side with the better one marked on every row.
    • It starts before the answer. The moment you send, the panel opens with the skill's real shape, empty and honest: the headings, the rows, the controls, and one line on what is coming. Nothing is invented; the only content is your own words. Ask about a mortgage of 250k at 4.5 percent over 25 years with a 30k deposit and the calculator opens with those numbers already in it, free, while the model is still thinking. Say what is broken and the Fix card appears with its steps to come. Ask a question and the pages show up as askr reads them. The ghosts shimmer while a reply streams and hold still when it ends, and on a phone a slim bar appears at send, pulses while askr works, and opens the sheet.
    • Price, Balanced or Performance. Sits in the message box where the mode chip was, beside the model it runs. Price runs a cheaper model with short answers, Balanced is the everyday middle, Performance runs the strongest models, asks for thorough checked answers, and in Health, Money and Fix it also searches the web. Tapping a tier switches the model; picking a model by hand moves the tier to match, or to Custom. The choice is remembered for that skill's next chat. There are no cost estimates inside a skill: the receipt under each reply says what the turn cost.
    • Your settings as pills. Each skill's own settings sit over the message box as one kind of pill: your device and its system, the languages and the tone, the level and whether askr guides or explains, your country and currency, what you need from a purchase, the length and format of a summary. Menus and panels float over the chat instead of pushing it down, and the settings carry to your next chat in that skill.
    • Free on your device. The calculators open free, with example numbers. The arithmetic check in Learn, the quiz marking, the word diff in Write, the lens in Brainstorm and the reading level all run on your device, with no model call and nothing charged.
    • Double-check with another model. In Look things up, Learn, Health, Money and Summarise, one tap asks a model from a different family to look for mistakes in the last answer and say what it would change. Only when you tap it, priced as one turn on that model.
    • Photo first. Fix it, Health, Translate, Money and Learn put a camera button first: a photo of the label, the report, the menu, the statement or the question. Summarise and Health also take a PDF or a document.
    • Save or share. Whatever the skill made, the draft, the steps, the appointment list, the shortlist, the comparison, the sources, copies as text, saves as a file, or goes to your phone's share sheet.
    • A skill offered from plain chat. When a plain chat asks about a loan, a symptom, a translation or a fix, a chip offers to ask it in that skill with the question ready to send. Decided on your device from your words, no model call, easy to dismiss.

    Improvements

    • How it works. Each skill's instructions are held on our server, added to every turn of that chat with the date in your time zone, your language and your settings, and no client can write them. The model answers in plain words plus fenced blocks of structured data with a kind, such as steps, draft, labs, calculator or citations. Your device cuts the blocks out of the reply, draws each with the skill's own component inside its own error boundary, so a bad block costs one card and never the chat, and shows the rest as ordinary markdown, with maths typeset.
    • What is checked, and where. A citation in Look things up must name a page the search actually read for that answer; one that does not is repaired to the real page behind the same number, without its quote, or dropped. A quote in Summarise must be words from the text you pasted. The maths in Money, the arithmetic in Learn and the diff in Write run on your device, never in the model. If your words describe an emergency, Health shows the care card on your device before the reply and tells the model to agree with it. Nothing is trusted just because a model said it.
    • Live reference data. Health reads the matching MedlinePlus topic and Money the Bank of England Bank Rate, UK inflation, exchange rates and, when you ask, crypto prices, fetched before the answer, cached, with a short deadline, so a slow source never holds a turn up. Fetched text is wrapped as data the model is told to quote, not obey.
    • Every model, one contract. Because the blocks are plain fenced JSON, every model on askr can run every skill; each skill simply opens on the model that suits it, a reasoning model for Learn and Money, a fast one for Translate and Summarise, the strongest for Performance.
    • One size everywhere. The skill header is one slim row in every skill, the settings are 30 pixel pills, every label is in the site's own face and sentence case, and everything a skill adds over the message box sits on the message box's own column at any screen width.
    • Search turns always answer. Before the last call of a search turn, the one with no tools, the model is told in plain words that no more searches can run and to answer from what it has. Without it, one turn in fifteen on a fast model asked for another search and wrote nothing. No extra call, no extra cost, and plain chats get it too.
    • The message box. The Compare models side by side button is gone from the message box; the Compare page itself remains. The cost of a turn shows once, beside send, or in the model pill on a phone.

    Bug fixes

    • A shared chat link opened on the home. A link with a chat in its address, opened in a fresh browser, landed on the home page instead of the chat. The address was read after the first network wait, by which time it had already been cleared.
    • Reopened search chats. A chat that searches now shows the globe lit when reopened, as it did when it was started.

    Also

    • Prompt tests. All ten skills were run at all three tiers against a judge that sees the history, the settings and the sources, and every block those runs produced was drawn through the real screens on a server. The test scripts ship with the code.
    • Testing on staging. Hannah built the ten skills on an isolated branch and tested them on staging; the polish, the frames and the release went through staging before production the same day.

    Finished: askr Skill

    The announcement on X
43. 1.395Time to think24 September 2026The Claude 5 family thinks before it writes, and on askr that thinking used to arrive as silence: the chat gave up after three minutes, then the connection upstream gave up at five.

    Claude now thinks where askr can see it, the chat says what is happening while you wait, and a brief Claude cannot finish is handed to a model that can.

    New features

    • Claude thinks out loud. Claude Opus 5.5, Sonnet 5, Fable 5.1 and Haiku 4.5 used to think in silence: not a byte arrived until the answer began, and after about five minutes of that the connection was closed upstream (code W511). Every Claude call now takes the route that streams the thinking as it happens, so the chat shows thinking from the first seconds and a long think is no longer cut short.
    • Said before it happens. With a Claude model picked, the composer says it thinks before it writes and a big brief can take several minutes. Once a Claude reply has been thinking for 45 seconds, the card says so in full: that it may run out of room, that nothing is charged if it does, and where the brief goes next.
    • Handed on, not dropped. A reply, thinking included, may be up to 32,768 tokens. When Claude thinks through that whole budget and never writes (code W514), nothing is charged and the same brief runs again on GPT-6 Sol as its own answer, under its own name, without you asking. A direct line to Anthropic is being built into our router so Claude gets the room it needs.

    Improvements

    • Ten minutes to think. A model has ten minutes before its first word and three minutes between words once it is writing; an answer may run for twenty minutes in all.
    • With Search the web too. The wait after the searches, where the model reads the results and thinks, has the same limits and the same notes.

    Bug fixes

    • Stopped after one sentence. With Search the web on, three minutes of quiet was taken for a dead connection and the answer stopped with whatever had arrived. Those turns were charged only for the sentence that arrived, and where an allowance covered them, nothing came off the balance.

    Announcement to follow.
44. 1.39File attachments24 September 2026Give any model a file.

    Pictures, PDFs, text and code, Word, PowerPoint and Excel, audio and video go in from the paperclip, in the app and in Collab rooms. A model that cannot read a file itself gets its words, so a PDF works on every model.

    New features

    • Attach almost anything. The paperclip, a drop or a paste takes pictures (PNG, JPEG, WebP, GIF, up to 8 MB), PDFs (32 MB), text and code files (5 MB), Word, PowerPoint and Excel documents (32 MB), audio (MP3, WAV, M4A, OGG, FLAC, 25 MB or 20 minutes) and video (MP4, MOV, WebM, 100 MB or 10 minutes). Up to 6 pictures, 3 PDFs, 5 text files, 3 documents, 2 audio files and 1 video on a message. Uploading is free; a file costs tokens only on the messages that carry it, and the estimate and the receipt both say so.
    • PDFs and documents work on every model. A model that reads files gets the file. Every other model gets its words, read once on the way in, so the cheapest model in the picker can read your contract too. A PDF over 100 pages goes as text everywhere.
    • Audio and video, on the models that take them. 32 chat models hear sound and 66 watch video, as the catalogue stood on 24 September. Audio for any other model goes as a transcript, made once and kept with the file.
    • A Files page. Everything you have uploaded, for chats and Collab rooms, in one list: what it is, what was read from it, and when. Open, download or delete. A chat's files go when the chat goes; an upload never sent is cleared after seven days.
    • In Collab rooms. Attach the same kinds in a room. The file is the room's, readable by its members, and stays with the room.

    Improvements

    • Pictures at full size for edits and video. A picture for an image model to change, or the start frame of a clip, now goes at up to 4,096 pixels on its longest side and 6 MB, not shrunk for chat. 212 chat models see pictures.
    • Ask again about a file. Your three newest messages that carry files send them in full; older ones are named in a line, so a PDF from the start of a long chat is not paid for on every reply. Files count toward the model's window, and a turn that does not fit says so before it runs.

    Bug fixes

    Nothing this time.

    Also

    • In the docs. Attach files has the limits, the per-model rules and the costs; codes W604 to W612 on Error codes say what a refused file means.

    Announcement to follow.
45. 1.388Credits and clips24 September 2026Airdropped credits are spent first, then credits given to you, then the credits you bought, and every receipt says which.

    Runway Gen-4 no longer offers a 10-second clip the gateway does not deliver, a clip shorter than the one you paid for is charged at the shorter price, and the render confirmation is askr's own.

    New features

    Nothing this time.

    Improvements

    • Spend order, spelled out. Free (airdropped) credits go first, then credits given to you (a referral bonus, a transfer in, an operator's credit), then the credits you bought. Every receipt and the balance panel say which pot a run came from.
    • Charged for the clip that arrived. Every finished clip is measured before it is charged. Shorter than the length you paid for, on a model that prices the shorter length, and you pay the shorter price. askr eats the difference.

    Bug fixes

    • Runway Gen-4 renders 5 seconds only, for now. The gateway bills the 10-second price and returns a 5-second clip: on 24 September three request shapes all came back at 5.04 seconds. The 10-second option is withdrawn until it works, and everyone charged for it is refunded.
    • Render it? is askr's own dialog. The browser's confirm box could be silenced by the browser after a few renders in a row, and the Render button then looked dead until a reload. It cannot happen with ours. The same for the prompt when you edit a message that has replies.

    Announcement to follow.
46. 1.387Scroll while writing24 September 2026Scrolling up during a fast answer no longer pulls you back to the bottom.

    The chat follows an answer until you take hold of it, and follows again when you scroll back down.

    New features

    Nothing this time.

    Improvements

    Nothing this time.

    Bug fixes

    • Take hold and it stays. Since 1.383 the chat followed an answer only while you were near the bottom, but a fast model wrote quicker than you could scroll out of that zone, so a flick was undone within a token or two. The chat now stops following the moment you scroll up, however fast the answer arrives, and the answer keeps growing below you until you come back for it.
    • On a phone too. A flick that left the transcript still moving when your finger lifted used to count as being at the bottom, and the next token dragged you back. The direction you moved decides now, not where you happened to be when you let go.
    • Back down, it follows again. Scroll to the bottom and the chat follows the answer as before. A new message of yours still goes to the bottom, and so does opening another chat.

    Announcement to follow.
47. 1.386Free credits first24 September 2026Airdropped credits are spent before your own balance, on every model.

    The balance panel in chat now shows exactly that: an answer paid from free credits moves the free line, not your balance.

    New features

    Nothing this time.

    Improvements

    • Spent this session counts both. The line under the balance adds what the free credits covered to what the balance paid, so it is what the session cost, not only the part you paid for.

    Bug fixes

    • Free credits go first, on every model. A turn, a picture or a clip comes out of your airdropped credits until they are gone, and only then out of your balance. That has been the rule since the first airdrop; the panel in chat was the part that got it wrong.
    • The panel moves the right number. It used to take every charge off your balance on screen and leave the free line alone until you reloaded, so a session on free credits looked like it was spending your own money. Every receipt now says how much the free credits covered, and the panel takes that off the free line and only the rest off the balance.
    • Read back after anything irregular. A stop, a reload mid-answer or a render checked later re-reads both numbers from the account instead of guessing.

    Announcement to follow.
48. 1.384The whole window24 September 2026Every chat on askr, in the workspace and in collabs, now carries up to 25 times more per turn: up to about 3 million characters, sized to each model's own context window, which is about 1 million tokens on the largest models.

    New features

    • 25 times more per turn. askr used to send any model at most 120,000 characters of a conversation, about 40,000 tokens, however much the model could hold. Each turn is now sized to the model's own context window, with room kept for the reply, up to about 3 million characters. 351 of askr's chat models get bigger turns.
    • Up to 1 million tokens. The largest models, Kimi K3, GPT-6 Sol, Claude Sonnet 5 and DeepSeek V4 among them, can now use their whole window of about 1 million tokens. A model with a 200,000-token window goes from about 40,000 tokens a turn to about 167,000.
    • Whole documents, one question. One message can now fill the whole turn, where it used to stop at 48,000 characters. Paste a codebase, a book or a stack of reports and ask about all of it at once.

    Improvements

    • Long chats keep their start. Up to the last 400 messages are sent to the model, up from 20, so a long conversation no longer forgets how it began.
    • In collabs too. A room now reads up to 400 messages of its thread and as much as its model's window holds, up from 20 messages and 100,000 characters, and one message can run past the old 24,000 characters.
    • Kept whole. A long chat is saved in full to your account and stays in your history, without pushing your other chats out of the browser.
    • You pay for what you send. Tokens are priced as before, and a bigger turn holds more credits before it runs. A model with a small window keeps the limits it had.

    Bug fixes

    Nothing this time.

    Also

    • Tested at full size. A 779,000-character document, about 161,000 tokens and six times the old limit, went to Kimi K3 in one message. The answer came back in 20 seconds, with a detail from the very first line.
    • In the docs. Workspace limits and code W106 now say so.

    Announcement to follow.
49. 1.383Room to think23 September 2026Long code answers and big build briefs no longer run out of room.

    The reply ceiling is four times larger, a model that spends it all thinking now says so, a long answer is given the time it needs, and the chat no longer drags you down while it writes.

    New features

    Nothing this time.

    Improvements

    • Four times the room. A reply may now run to 32,768 tokens, thinking included, up from 8,192. Long code answers arrive whole instead of stopping mid-file. A thin balance gets the largest reply it can hold rather than a refusal, and a model that takes less than the ceiling is retried one size down on its own, so nobody sees a refusal for it. If an answer still reaches the ceiling, the receipt says so and "continue" carries on.
    • Time for a long answer. The four-minute cap on an answer is gone. As long as the model keeps writing, the answer keeps coming, for up to twenty minutes, so a long file is no longer cut for time either.

    Bug fixes

    • A model that thinks itself out of room says so. Some models, the Claude 5 family among them, think before they write. When it used up the whole budget the reply said "did not answer, try again", which sent people round in circles. It now says what happened, with code W514, and suggests a model that thinks less. Credits go back either way.
    • Read while it writes. The chat follows an answer only while you are at the bottom. Scroll up to re-read and it stays where you put it; scroll back down and it follows again. A new message of yours still goes to the bottom.

    Also

    • In the docs. The reply length and code W514 now say 32,768.

    Announcement to follow.
50. 1.382Send credits23 September 2026You can send credits to another askr account.

    Bought credits and credits you were given move; airdropped promo credits and the holder discount stay where they are.

    New features

    • Send credits from your balance. In Settings, Balance & deposits, a Send credits button beside Add credits: enter an @, an amount, a note, confirm, and both of you get a receipt and a line in your Sends list.
    • What moves and what does not. Credits you bought or were given move. Airdropped promo credits keep their expiry and never move, and the holder discount follows your wallet.
    • Limits that keep it safe. Minimum $0.10, credits deposited in the last 24 hours wait a day, and a six-digit email code for larger sends.
    • Final, with a safety net. A send is final once confirmed. If one was made by mistake or by someone who was not you, tell us and we can reverse it while the credits are still there.

    Improvements

    Nothing this time.

    Bug fixes

    Nothing this time.

    Also

    • Still spend only. Sent credits buy model usage like any other; nothing here turns credits back into cryptocurrency or cash. The terms say so. See Send credits.

    Announcement to follow.
51. 1.381Search the web23 September 2026Any model can now search the web and read pages when it needs live facts.

    Off by default; a globe in the composer turns it on, and every search is priced before you send.

    New features

    • A globe beside attach. The model may search when your question needs something current. The globe glows while a search runs and shows how many ran. Remembered per chat, with a setting to have it on for every new chat.
    • Priced before you send. The estimate adds the most a turn can spend on searching, half a cent a search plus the page text at the model's rate. The receipt shows what actually ran, for example 2 searches, $0.01. At most 3 searches and 2 pages a turn.
    • Sources on every answer. Numbered [n] marks in the reply link to a Sources line underneath, so you can check what the model read.
    • It knows when to look. Ask about today, a price or the latest news with the globe off and the search switches itself on for that chat, the globe glows, and the receipt shows what it cost. Turn the globe off in a chat and it stays off there.
    • In Collab rooms too. A globe in the room composer, per room. An @askr line that needs live facts searches, the globe glows for everyone in the room, the reply carries its sources, and the searches come off the room balance like the rest of a room's model use.

    Improvements

    • Only where it works. The globe is greyed out on models that cannot use tools, with a note saying so. Zero-retention versions of a model search too.

    Bug fixes

    Nothing this time.

    Also

    • Privacy: with Search the web on, the search terms the model chooses go to Brave Search; pages are read by askr's server, never from your browser. See Search the web.

    Announcement to follow.
52. 1.38New chat UI23 September 2026The chat, the collab rooms and everything around them get a new look: the moon and its falling light, a rail that opens into a sidebar, glass surfaces, Geist type, and a light mode beside the dark one.

    Every result is still priced before you send it.

    New features

    • A new look for the whole app. The chat at /chat, every collab room and the pages around them are redrawn: the red moon with its falling light lines on the home, a 64px icon rail that opens into a 264px sidebar, glass chips for the $ASKR price, your balance and your account, and Geist type with every number, price and label in Geist Mono. On a phone: a top bar, a menu drawer, and a composer docked above the keyboard.
    • Light mode. Choose Light, Dark or System from the sun/moon switch in the rail, the sidebar, the phone menu and the Collab room, from your account menu, or in Settings under Appearance. The app opens in Dark unless you choose otherwise, your choice never flashes the wrong theme, and the site and the docs stay dark.
    • A home that gets out of the way. The askr logo, then one line, "What are we building?", whose last word rolls through more than seventy words. Five mode tiles in one row: Ask, Code, Images, Video and Collab. The composer is glass, grows as you type, shows the model and its price before you send, and takes attachments, compare and dictation. Recent chats and your Library sit below it. The Audio and Skills tiles, the old welcome line and the footer line are gone.
    • The model picker. Cmd-K opens a dialog on desktop and a bottom sheet on a phone: search, the pills All, Ask, Code, Image and Video, A to Z and price toggles, keyboard from top to bottom. All lists every model on the platform, Ask the popular and cheap general ones, and Code every text model that can write code, with the code-flavoured ones first.
    • Add credits without leaving the chat. Add credits opens as a dialog or a sheet from the balance chip, the sidebar card and the phone balance pill: pick the asset from the ones the processor offers, the amount, and you get the address and a QR that scans in both themes. A deposit you start and close is watched to its end and lands on your balance.
    • History, Search, Library, Collab and Settings as pages of their own. History groups your chats by day with mode filters, what each chat cost, Rename and Delete. Search covers your chats and your library with highlights. Library shows your pictures and clips as tiles with Download, Re-run, Open chat and Delete. Settings holds your account: General with Appearance and a default model per mode, Balance and deposits, $ASKR benefits and Privacy. /account now lives at /chat/settings.
    • Edit forks the chat. Edit a prompt you sent and it goes out again from that point as a new, charged turn, while the old words and the replies they got stay as an earlier version. A control beside the turn flips between versions.
    • Collab rooms. The room title is a dropdown with a Switch room popover, the chat column sits beside cards for People, Threads and Room settings, and a phone gets sheets for each. Every member sees the room settings and the room name; only the creator changes them. Reactions sit on the bubble they belong to. A room that cannot load says so and offers Try again. "Collab rooms" is now just "Collab".

    Improvements

    • Calm inside a chat. Once a conversation starts, in any mode, in a collab room and on the pages, the light lines stop and the moon dims, so nothing moves behind what you are reading. The home keeps its motion. Reduced-motion settings are respected everywhere.
    • Replies with receipts. Every answer carries the askr mark, the model, the mode and its status, and a receipt line that starts with the price: Copy, Read aloud, Re-run, Compare and Download as plain links. Code blocks have a header with the file name, diff counts and tinted added and removed lines. Pictures show two up, compare shows four cells with Pick, clips show on a card with a rendering row while they render, and Read aloud plays in a small player that fits a phone.
    • Chats run on their own. Each open chat keeps its address, like /chat?chat=abc, so a reload or a link comes back to it, and /chat on its own is the home. A reply still writing in one chat no longer blocks sending in another, and Stop only stops the chat on screen. Home always opens a fresh chat.
    • Everything from 1.37, in the new look. Failure cards that say why and whether credits went back, pictures and clips shown whole at their own shape, one-click copy, editing a sent message and deleting from your library are all carried into the new chat and collab rooms, and a reader at the foot of a chat stays there when a picture loads.
    • Built to be used with a keyboard and a screen reader. Every control is reachable by Tab with a visible red focus ring, dialogs trap focus and close on Escape, the phone drawer makes the page behind it inert, fields show focus on their rounded box, hit areas are at least 44px on phones, and text meets WCAG AA in both themes.

    Bug fixes

    Nothing this time.

    Announcement to follow.
53. 1.375Second supplier23 September 2026Many models are now served by a second supplier, Engy, at a lower price and with zero data retention.

    Pick the version you want; nothing else about how askr works changes.

    New features

    • The same model, cheaper. In the model picker a model served by both suppliers shows a second row beneath it, "by Engy", with its own price. Pick either.
    • Eight models to start. The saving is against the same model on askr before, measured on a standard turn. DeepSeek by Engy: V4.1 Flash 86% less, V4 Flash 84% less. GLM by Engy: 5.2 62% less, 5.3 34% less, 5.3 Flash 14% less. Kimi by Engy: K3 38% less. Qwen by Engy: 3.8 27B 90% less, 3.6 35B 71% less.
    • Zero retention on those rows. Every Engy version except one is served on hardware Engy operates, with prompts and answers never stored or trained on. The row carries a ZERO RETENTION chip and the model button says so.

    Improvements

    • Priced the same way as everything else. The estimate beside Send, the receipt after a reply and the holder discount all work as before; the receipt says BY ENGY when that version ran.

    Bug fixes

    Nothing this time.

    Moved on: Router rebuild

    Announcement to follow.
54. 1.37Feedback fixes23 September 2026Fixes and improvements from what people told us.

    When a model says no, askr now tells you why and that you were not charged, instead of a failure that looks broken.

    New features

    • One-click copy on everything. Every prompt and every answer has a Copy, in the workspace and in a collab room: words copy as text, a picture copies as the picture itself where your browser allows, and a clip copies its link.
    • Edit what you sent. In the workspace, edit a prompt and send it again from that point, after a check that says how many messages it replaces; what those messages made stays in your library. In a collab room, fix the words of your own message, and everyone sees it marked edited.
    • Delete from your library. Every picture and clip in your library now has a Delete that asks first, so nothing goes with one stray click. A deleted item is gone at once; only the billing record of what it cost stays.
    • Live prices on the homepage. The card on the homepage now shows the newest Claude, the newest ChatGPT and the most used model from the next most popular brand, each with its price and the $ASKR holder price, which applies within your daily allowance.

    Improvements

    • Pictures and videos shown whole. In a chat or a collab room, a generated picture or clip now shows its whole frame at its own shape, sized to fit, instead of a square crop until you click it. Reading further up while one loads no longer pulls you to the bottom.
    • The same message every time. You see the same reason whether you wait on the page or come back later, and never the provider's raw text.

    Bug fixes

    • When a model says no, you hear why. A request a provider refuses under its content rules, like a video of a real public figure, now says so in plain words, tells you nothing was charged, and suggests rewording it or picking another model. Codes W816 for video, W415 for chat and images.
    • No more "try again" for something that cannot work. A refusal never tells you to retry unchanged, because the same prompt would be refused again. Rules differ by provider, so another model may accept it.
    • Videos no longer hang on a refusal. A render the provider blocks is marked failed as soon as the provider says so, instead of sitting on "taking longer than expected", and a brief network blip while we check on a render no longer fails one that is still going.
    • A failed job looks like a failure. In the workspace, a request that fails now shows a failure card with the reason, and says whether your credits went back or nothing was charged, instead of reading like an answer.
    • Long videos keep their place. A clip still rendering after 7.5 minutes, or through a dropped connection, keeps being watched and lands on its card when it finishes, even after a reload.
    • In a collab, a clip that fails says so, and its reserved credits go back to the room, and a refusal never tells the room to try again.
    • Footer: the chain badge, email and copyright now line up on one edge, on desktop and on phones.

    Also

    • Every code explained: workspace error codes now lists W415 and W815 to W819.

    Announcement to follow.
55. 1.36ASKR Connect22 September 2026Link the wallet that holds your ASKR without connecting it, signing anything or sending a single token.

    Paste the address, buy a little ASKR from that wallet in the app you already use, and the chain shows it is yours. Built for FOMO and every app that can buy but cannot connect.

    New features

    • Two ways in. Connect wallet signs a message, as before. Verify holdings is new, and needs nothing but your address.
    • No connection, no keys, nothing sent. Paste your address, press start, and buy any amount up to 25,000 ASKR from that wallet within sixty seconds. Tap Done any time in the next ten minutes and the wallet is linked.
    • Works from FOMO. A FOMO buy is made by your own FOMO wallet, so it proves the wallet is yours. Not sure of your address? The page shows you where it is: in FOMO, Account, then Deposit, then Deposit crypto.
    • Only you can do it. Anyone can send tokens to an address, so a balance going up proves nothing; a buy the wallet made itself does. Tokens someone else sends you never count, and neither does a buy made before you pressed start.
    • The same benefits either way. A wallet linked by a buy gets the same discount and the same access as a signed one, with its first allowance at the next 00:00 UTC.

    Improvements

    • Holdings stay fresh. A linked wallet's balance is re-read within ten minutes of any ASKR moving in or out of it. The daily allowance is still set once, at 00:00 UTC.

    Bug fixes

    • pump.fun and other apps that buy through a third-party router. A buy that reaches your wallet through a router cannot show the wallet is yours, so the page now asks you to sell the ASKR you just bought, and that sale confirms it.

    Also

    • Both ways, step by step: Holding ASKR.

    Moved on: Holder access

    Announcement to follow.
56. 1.35Holder discount21 September 2026Hold ASKR, link the wallet, and every model on askr costs you a twentieth of its price on the credits you bought.

    How much you get at that price each day scales with how much you hold: $1 of AI a day for every 0.01% of supply, capped at $25. It resets at 00:00 UTC.

    New features

    • Every model at 5% of its price, not a fee discount: askr is feeless, so the 95% comes off what the model itself costs and askr covers the rest.
    • On the credits you bought. The holder price applies to credits you deposited, and the allowance counts only those. Airdropped credits and credits given to you are spent first, at the ordinary price, so a free credit is never doubled by the discount.
    • Your allowance scales with your holding: $1 of AI a day for every 0.01% of supply you hold, capped at $25 a day. Past the allowance you pay the ordinary price, so nothing ever costs you more than it would without the token.
    • Read once a day, at 00:00 UTC. At that instant a bag of tokens sits in exactly one wallet, so it earns exactly one allowance wherever it moves afterwards. Splitting a holding across wallets earns the same as holding it whole.
    • A wallet you link today earns its first allowance at the next reset, which is what stops anyone linking, spending, moving the tokens on and linking again.
    • Unused allowance does not roll over. It is the size of the discounted lane for the day, not a grant.
    • Shown everywhere money is: on your balance, in the workspace rail, on the wallet panel and inside a collab room, with the discounted credits left today and the time it resets. It shows at 0 balance too, so you can see what a deposit would buy.
    • It works inside a collab too: a run in a room is charged against the runner's own allowance first, and the room's purse pays whatever is left. An allowance is per person per day either way, so a room of holders is simply cheaper to run.
    • On the price list: every rate on /pricing now carries the holder price beside it.

    Improvements

    Nothing this time.

    Bug fixes

    Nothing this time.

    Also

    • You still need credits. The discount makes the AI cheap, not free: a balance is what the 5% is charged against, and a deposit starts at $5.
    • The docs say the real rule: Holding ASKR replaces the old promise of 25% off a platform fee that was never charged.

    Moved on: Holder access

    The announcement on X
57. 1.31Collab fixes21 September 2026What the first rooms needed: pictures and replies and reactions inside a room, code that reads the same in it as out of it, and the joining bug that stopped some invites working at all.

    New features

    • Talk as well as prompt: attach pictures, reply to a particular message, and react with an emoji. The picker opens where you pressed it, not in the corner.
    • Code reads the same everywhere: coloured, labelled by language, one click to copy, inside a room exactly as outside it.

    Improvements

    • Your @ is where your picture is, on the profile, instead of buried in the referrals card.

    Bug fixes

    • Join actually joins. An invite link could land you on a room you were not in. Fixed the day it was found.
    • The privacy page says what a room actually does with what is written in it, and what is forwarded to a model. It said three things that were not true of collab.
    • Crediting an account with no email works again in the admin. Accounts have been able to exist without one since 1.1, and the action still assumed every account had one.

    Also

    • Page views, counted: analytics on every page, named in the privacy notice and the policy rather than added quietly.

    Announcement to follow.
58. 1.3askr Collab20 September 2026An invite-only room where a few people work the same models together out of one purse.

    Chat, code, image and video, the same four doors as the workspace, with everyone in the room seeing every answer as it arrives.

    New features

    • A room with its own purse. Whoever opens it funds it, and every run in the room is paid by the room. Nobody spends their own balance by being invited.
    • Invite-only, by link. The room is not listed and not searchable; the link is the door.
    • All four doors inside it: chat, code, image and video, on the same models as the workspace.

    Improvements

    Nothing this time.

    Bug fixes

    Nothing this time.

    Also

    • A beta, and it says so. Rough edges are known and being worked; tell us what breaks from the box on the roadmap.

    Announcement to follow.
59. 1.25Voice and roadmap20 September 2026Speak to askr on any device and it goes straight to the model you chose; every reply can read itself aloud.

    Both run in your browser, cost nothing and send no audio to askr. The roadmap now lives on the site, with a box to ask for what you want built.

    New features

    • Dictation: a mic beside the paperclip in the composer. Tap to listen, tap to stop; sending stops it. Words show as they are guessed and settle as you go, and a pause to think does not end the session.
    • Read aloud: a link on every finished reply, in your browser's own voice. Code is skipped with a word, links are read as their text, long replies are read in pieces so it finishes. One reading at a time; tap to stop.
    • See the mic hearing you: four bars beside the mic move with your voice while it listens.
    • The roadmap, on the site: heyaskr.ai/roadmap shows every version shipped and the plan ahead in order, and has a box that sends what you want built straight to the team.

    Improvements

    • The composer fits what you type: the box grows and shrinks with the message, typed or dictated. On a phone the text runs full width with the buttons underneath.
    • The site asks for the microphone on the first tap, and if you have said no it tells you where the switch is on your platform.

    Bug fixes

    • Docs put right: deposits are credited in full (askr absorbs the processor's cut), overpayments are credited, chats live on your account, and referrals have their own page.

    Also

    • In your browser, not on our servers: Chrome, Edge and Safari carry the speech engine, so the buttons appear there and stay hidden in Firefox. No audio is sent to askr, nothing is recorded, and neither feature spends credits. The privacy page says so.

    The announcement on X
60. 1.2Handles and referrals20 September 2026Every account gets an @ and a referral link.

    Anyone who joins through it pays you for life: 10% of their top-ups in credits the moment they land, and 10% of their fees once fees are on. No ceiling.

    New features

    • An @ on every profile: a username account's handle is its @; email and wallet accounts can claim one. heyaskr.ai/r/<your code> is your link. How referrals pay.
    • 10% of every referred top-up, paid to you in credits the moment their deposit credits. Automatic, once per deposit, never out of their balance.
    • 10% of their fees, when fees are on. Accrued per run and paid hourly as credits. Nothing accrues while v1 is feeless; switching the fee on is the only change left for that to start.
    • A referral link survives the sign-up: the code follows every door (username, wallet, email) and holds for 30 days.
    • Support where things go wrong: @askrsupportbot in the chat rail, under any refunded reply, on the top-up card, in the account tabs and at the foot of sign-in.

    Improvements

    Nothing this time.

    Bug fixes

    • Wallet fixes: a wallet that is an account's only way in can no longer be unlinked; BNB pays from a wallet with a proper QR; WalletConnect opens one modal, not two; Robinhood Chain is the active chain on phones.

    The announcement on X
61. 1.1Made for crypto19 September 2026Sign in with a username and password or with a wallet, no email needed.

    Link the wallet that holds ASKR and the account shows what you hold. Add credits the crypto way.

    New features

    • Sign in without an email: a username and password, or a wallet signature. An email can be added later for recovery.
    • Wallet sign-in on any wallet, through WalletConnect as well as browser wallets.
    • Holdings on your account: ASKR held, share of supply, your band and the fee discount it earns, read from the chain on link, on demand and daily.
    • Add credits the crypto way: ten coins as tiles with a search under them, a checkout card with the address in a box and a Copy button, and a tracker that follows the deposit from sent to credited.
    • A receipt by email when a deposit credits.
    • Support on Telegram: @askrsupportbot takes a ticket and relays the answer.
    • Chats follow the account across devices; the history is named by what each chat asks.
    • Drop a picture anywhere on the page to attach it; tap a generated image to see it full screen.
    • The code panel lists files by name, one open at a time, with Save all.
    • The token page carries the contract address, the Team Finance lock, and the Buy and Chart buttons.

    Improvements

    • Models that stop answering hide themselves from the picker instead of failing your request.

    Bug fixes

    Nothing this time.

    Finished: Holder access

    The announcement on X
62. 1.0askr is live19 September 2026, where we startedheyaskr.ai is open: every major model behind one sign-in, priced at upstream cost, paid in crypto with no card and no subscription.

    ASKR launched on Robinhood Chain the day before.

    New features

    • Four doors from one balance: chat, code, image and video.
    • Feeless v1: a dollar in is a dollar of model; the processor's and the gateway's fees are absorbed.
    • ASKR on Robinhood Chain, launched 18 September, an access token with bands by share of supply.
    • Docs at home on heyaskr.ai/docs, generated from the code, reading the catalog live.

    Improvements

    Nothing this time.

    Bug fixes

    Nothing this time.

    Finished: LaunchMoved on: Holder access

    The announcement on X

1. Step 1Our own payment processorIn progressDeposits without a third party in the middle.

   Any token on any chain we support lands straight where credits are honoured: no processor fee to absorb, no forwarding step that can fail, more chains and coins added as fast as we can watch them, and settlement we control end to end.

2. Step 2Router rebuildIn progressaskr is moving its routing onto the model makers' own APIs, one maker at a time.

   Claude, GPT and Grok already run on Anthropic's, OpenAI's and xAI's APIs as the versions marked (direct), at the maker's list price, with the model's thinking shown as it happens and the conversation cached between turns, and Seedance's (direct) versions run on ByteDance's own API at less than half the price. More makers follow the same way. Cheaper per request, faster, ours to control.

3. Step 3askr AnonIn progressPrivate chats, when you need them, on top of askr's standard privacy.

   Live since 1.78: no account, every chat locked with its own key pair made in your browser, sealed models first (GLM-5.3, Kimi K3, DeepSeek V4.1 Flash and more, running inside secure hardware that proves what it runs), zero retention after that, a warning before any model that keeps your words, one random id per chat for model companies, nothing else reading it, and everything erased when the chat ends. Next: end to end on sealed models, so askr only relays scrambled words and never holds a readable copy; pictures, clips and files locked like messages; no IP addresses kept, and a Tor address; zero-knowledge payment proofs and prepaid codes, so a payment can't be tied to a chat; and open code, a public record of the running build, askr's relay in sealed hardware and an outside audit, so anyone can check it.

4. Step 4askr AnywhereIn progress@askr where your group already talks.

   Live since 1.92 on Telegram and Discord: add askr to a group or a server from any chat, tag it there for anything askr does, and every question and answer is kept in a room of that chat, paid by the chat, with who can ask set by whoever started it. X and WhatsApp are not in it yet.

5. Step 5askr ModelsPlannedTokenised official AI models airdropped to $ASKR holders, with benefits of free usage, instant sales and LP rewards.

   Each model token is paired with ASKR.

6. Step 6Chev buys thingsPlannedChev fills a checkout itself, so a card means something: your own Visa through Link, with a yes on your phone for each purchase; a Visa on Chev's wallet that spends its dollar coins; or a card you give it.

   Every purchase under the wallet's modes and ceilings, and every one listed with what, where and how much. The cards are built and off; the checkout comes next, and we knock these out as we go.

7. Step 7Chev everywherePlannedChev beyond Telegram: direct messages on Discord, an email address for each Chev that only its owner may write to, with the answer a reply in the thread, and one iMessage line to text.

   Each switches on as its account opens, and whatever you say on any of them shows in the conversation in the app. Built in 1.82; knocked out one channel at a time.

Alongside every step: new models the day they land, new coins as they are supported, fixes every week, and changes from what people tell us.

The API at /v1 closes on the day we announce it. Credits keep working in askr.