How to use AI Chat Console

Sending your first message takes seconds — no setup required.

Suggestions for you: once you have used the app for a while, a new chat opens on a For you tab. It starts with the shortcuts you use most (they go into the message box, ready for your text) and the saved prompts you used recently, then adds ideas for the kind of work most of your chats are about — code, writing, learning, planning or creative work. Only you see it, and you can turn it off in Settings → Preferences → Suggestions for you.

Fill-in templates: the Fill-in tab on a new chat has short forms for common jobs — write an email, summarize text, plan a trip, a cover letter, a meeting agenda, a social media post, a product description, a reply to a review, explain a topic, compare options. Answer two to four questions and the request is written into the message box for you to check and send. A saved prompt with blanks such as {{topic}} asks for them the same way when you use it from the Prompt Library or For you; anything you leave empty is left out.

Help when you need it: ask “What can you do?” in a new chat (or use What can it do? under the suggestions, or Ctrl+K) and you get a list of everything switched on for your account, each with a Try it — answered for free, with Ask the AI instead if you wanted the model’s own answer. Now and then a short tip appears the first time it applies, such as after your first answer with code; Got it hides that tip for good and No more tips stops them all. What’s new sits at the top of the notifications tray (the bell above the chat list), with a dot on the bell until you have seen the latest change.

Different AI models are good at different things. Click the model name near the message box to open the picker and switch anytime — even mid-conversation.

• Everyday questions & writing — fast, low-cost models handle most tasks well and use the fewest credits.
• Complex reasoning, coding, analysis — premium models think more deeply (they cost more credits).
• Vision — models with a 👁 icon can look at images you upload.
• Web search — models with a 🌐 icon can pull in live information. Searches read the top pages in full, not just the snippets; ask to compare, analyse or research something and it goes deeper — more pages, and a second round for anything the first one missed. Click a source under the answer to see what it says and whether it was read in full. Choose Quick, Standard, Deep or Auto in Settings → Behaviour → Web search depth.
• Privacy — models with a 🛡 shield support Zero Data Retention.

Best for this message: once you have typed something or attached a file, the top of the picker suggests up to three models that fit it — one that can see your picture, a strong coding model for code or an error, one with a large memory for a long document, one that searches the web for something current, or one that thinks step by step — with the reason. Models that cannot read an attached picture are dimmed.

See every model before you start

• The Models page (Features → Models in the top menu) lists every model that is switched on right now, read live — nothing is typed in by hand.
• It is grouped by what the model makes — Chat & reasoning, Images, Video and Speech & audio — then by provider.
• Tags show what each one can do: Sees images, Reasoning, Web search, Deep research, Runs code, Transcription and Free.
• Type in its search box (“claude”, “vision”, “free”) to narrow the list. Some models are only in certain plans — the picker in the app shows exactly what yours includes.

Chat settings: the gear button beside Tools (on a phone, in the top bar; or Ctrl+Shift+,) opens one panel with everything that shapes this chat’s answers: the model, instructions for this chat, memory, answer length, tone and format, web search, reasoning and creativity. Reset to my defaults puts them all back.

Long pastes: paste more than about 2,000 characters — a report, a log, an article — and it is attached as a text file (Pasted text) instead of filling the message box. The AI reads all of it, and your chat shows a tidy file card. Press Paste as text on the message that appears to undo it, or turn it off in Settings → Preferences.

The buttons beside the message box add extra abilities to your chat. Anything you switch on beyond your usual setup — Reasoning, Deep research, web search on every message, Short or Detailed answers, a tone, a format, Focus — shows as a small chip above the message box, so you always know what will shape the next answer. Press ✕ on a chip to switch it back.

The Prompt Library

Not sure how to ask? Browse the Prompt Library — ready-made prompts with real examples of what each one produces. Open one, press Use in chat, and it lands in your message box with the right model already selected.

• Examples, not promises. A prompt can carry the picture, document, spreadsheet or audio clip it actually made, so you can see the result before you spend anything trying it.
• Save your own. Anything you keep retyping can become a prompt. Keep it private, share it with named colleagues, open it to everyone signed in here, or submit it to the public library.
• Fill-in-the-blanks. A prompt can declare {{placeholders}}; you are asked for each one when you use it, rather than having to edit the text by hand.
• Fork anything. Found one that is nearly right? Fork it, edit your copy, and the original is credited.
• Type # in the message box to pull up your prompts without leaving the chat.

Pointing at a file with @

Once you have uploaded something to a chat, type @ in the message box to list every file in that conversation — what is attached to the message you are writing, and everything uploaded on earlier turns. Pick one and its name drops into your message.

• It says which one you mean. With three screenshots attached, “@chart-2.png what is the trend here?” is unambiguous where “the second image” is a guess.
• It brings an older file back. Long conversations cannot carry every document in full, so older ones are summarised down to make room. Naming a file with @ pulls that one back in complete — including a picture from earlier in the chat, which the AI otherwise cannot look at a second time.
• Deleting the text undoes it. The mention is just the filename in your message; remove it and nothing extra is sent.
• It works right away. A file is nameable the moment you send the message you attached it to — you do not have to reload the conversation first.
• Files that have passed their retention window are listed but greyed out — the file itself is gone, so there is nothing left to bring back.

Pointing at a file in Draw, Video and Speech mode

@ works in every mode, not just chatting — but what it does depends on the mode, because each one can use a file for something different. You are only ever offered the kind of file that mode can actually use, so the list is never full of things that would do nothing.

• Reference images are nameable too. A picture you attach with the reference button in Draw or Video mode joins that conversation’s file list for the rest of your visit, so a second or third attempt can pull it back with @ instead of you finding it on your computer again.
• Speech mode: the mention marks the spot. “Here is this week’s update. @notes.md That is everything for now.” is spoken as your two sentences with the whole document read out in between. Put the mention on its own to have nothing but the document read.
• A document has to have readable text. If nothing could be pulled out of it — a scan with no text layer, for instance — Speech mode will say so rather than read out the filename.
• It is always optional. Every mode works exactly as before if you never type @.

Speech mode charges for the words it actually speaks, so a mentioned document counts towards the length limit and the cost of that clip — a long PDF is a long clip.

Bringing a large document or an image back into a reply uses more of the conversation’s context, and images cost more than text — so mention the file you actually need rather than all of them.

Memory carries useful facts about you from one chat to the next — your role, your preferences, how you like answers written — so you stop repeating yourself. You decide what it keeps, and you can switch it off for any single conversation.

• It learns your style — only with your OK. Ask for the same change three times in a week — Shorter, Simpler, More detailed, Bullet points, a more formal or friendlier tone — or rate three answers 👎 as too long or too short, and it asks once whether to remember that. Yes saves a memory you can see and delete here; No is never asked again.
• Two kinds of memory. Facts you add yourself, and facts learned automatically from your chats. Both are listed in Account → Memory, where you can edit or delete any of them.
• Three choices on every chat. Click the memory button beside the message box: Memory on (saved facts and earlier messages are used), Memory off (no saved facts are used or learned; earlier messages in the chat still are), or Independent (each message is answered on its own — no saved facts, no earlier messages, and nothing is learned). The choice belongs to that chat only.
• What new chats start as is your choice: memory on or off in Account → Memory, and “answer each message on its own” in Account → Preferences.
• Auto-memory can be turned off separately if you would rather add facts by hand. Recall still works; only the automatic learning stops.
• Recall depth is adjustable — send everything remembered, or only what is relevant to the current message. The trade-off is thoroughness against token cost.
• Turning memory off entirely, in Account → Preferences, leaves only Normal and Independent in the menu, since there are then no saved facts to switch.

Shortcuts are your own saved prompts — instead of retyping the same request every time (a writing style, a report format, a recurring image size), save it once and insert it in a couple of keystrokes.

• Shortcuts aren't just for text — they can also carry a saved image/video/audio configuration, so / can jump straight into Draw, Video or Audio mode with your usual settings pre-filled.
• Admins can publish preset shortcuts for everyone — those appear in the same / picker alongside your own, marked as presets.

The whole site now shares one design: the same buttons, dialogs, icons and spacing everywhere, in light, dark and OLED themes.

Illustration: the icon rail down the left, with Projects open full size beside it and the Ctrl+K search on top.

• Focus mode: press Ctrl+. (⌘+. on a Mac), click Focus at the bottom of the icon rail, or choose it in Ctrl+K. The chat list and the rail fold away so the conversation has the whole window; a small Leave focus button stays in the corner. It lasts until you leave it or close the tab, and never changes your saved sidebar setting.
• Easier reading: in Appearance, Reading width keeps long answers in a comfortable column on a wide screen, Line spacing adds room between lines, Reading font → Easy-read uses a plain, open font with a little more space between letters and words, and High contrast makes text and borders stronger.
• Icon rail: the column on the far left. Chats brings your chat list back; Docs, Notes, Files, Projects, Bots, Tasks, Assistant, Modules and Prompts open full size beside it and switch in place without reloading; Search opens Ctrl+K. Pick its colour (theme tint, solid, neutral or your own) or hide it under Settings → Preferences. On a phone the same shortcuts are in the + menu at the top.
• Reply actions: under every reply you see Copy, Retry and 👍 / 👎 — plus Canvas when the answer is long or mostly code. Quote, Pin, Save note, Different model, Speak, Print and Translate & listen are in More (…). Want one always in the row? Tick it under Settings → Preferences → Always show under replies (or use the link at the bottom of More).
• Code blocks: each one has Explain, Wrap, Copy and Save (as a file with the right extension); HTML, CSS and JavaScript also get Run, a safe preview. Explain is an ordinary message, charged like any other.
• Long messages: a very long message you sent (a pasted log or article) folds to about eight lines with Show all; the whole text is still sent and kept.
• While an answer is being written: Shorter, More detail and Change direction… appear under it. Each stops it (you pay only for what was written) and writes it again your way; the cut-off version stays one tap away with ‹ ›. If nothing has appeared after about 45 seconds, it says so and offers Try another model.
• Select text in an answer and a small menu appears beside it: Ask about this and Quote put just that part above the message box for your question; Explain simpler and Search the web ask right away (only when the message box is empty, so nothing you typed is sent by surprise); Save note keeps just that part in Notes; Copy copies it.
• Long answers: an answer with three or more sections shows Jump to chips at its top — one per heading. An answer of about 300 words or more also gets TL;DR, which writes a short summary above it (a small AI call, a few credits; the cost shows on the summary).
• Details: More → Details under a reply shows how it was made — the model (and any model that stepped in), what it used (web search, apps, skills, your knowledge, tools), the files on your question, its sources, whether your saved memory was included, how much was sent to the model and from where, and what it cost.
• Reshape: under the newest answer, More → Reshape rewrites it — Shorter, Simpler, More detailed, Bullet points, As a table, More formal or Friendlier. The rewrite replaces the answer as a new version; use ‹ › to flip back while the chat is open. It is charged like Retry.
• 👍 / 👎: rate any answer. A 👎 asks what went wrong (optional) and, on the newest reply, offers Try another model. Ratings are free, never sent to the AI, and help the site see which models do well. Turn them off in Settings → Preferences.
• Use as reference (right after Download on generated images and videos): in a normal chat it attaches the image to your next message; in Draw it becomes the reference; in Video it becomes the start image.
• Outline: in a chat with four or more questions, Outline on the line under the message box (or Alt+O) lists every question and your pinned messages; click one to jump straight to it.
• The tokens line above the message box shows the chat size, and in Draw, Video and Audio the estimated cost of the next send.
• Quick tour: four tips the first time you open the chat. Replay it any time from Settings → Preferences.

Open the site in your phone’s browser and choose Install or Add to Home screen. It then opens full screen like an app.

• Shortcuts: press and hold the app icon for New chat, Voice mode, Draw an image or Ask about a photo.
• Share into a chat: in any app, press Share and pick AI Chat Console. Photos, PDFs and documents are attached to a new chat; text and links go in the message box. (Installed app on Android; up to 10 files.)
• Press and hold a message for its menu: Copy, Share, Select text (turns text selection on for that message), and every action under the reply, such as Retry, Read aloud and Branch. More under a reply opens the same sheet.
• Swipe a message to the left and let go when the round quote mark fills in: the message is quoted in your reply (if you had selected part of it, just that part). Code blocks and tables still scroll sideways as usual.
• Camera: the camera button beside the paperclip opens the camera; the photo is attached to your next message.
• Hold to dictate: hold the mic, speak, and let go. A tap still starts and stops dictation as before.
• Jump between answers (off until you choose it in Settings → Preferences: on a computer only, on a phone only, or both): scroll and the ▲ / ▼ buttons appear beside the chat; they move one answer at a time. On a computer Alt+Shift+↑ / ↓ do the same. Separately, if a reply finishes while you are reading higher up, the scroll button says New reply.
• Swipe a chat to the right in the chat list for Pin, Archive and Delete (Delete can be undone for a few seconds).
• Model list: slides up from the bottom. Use All, Recent, Favourites (tap ☆ beside a model) and Free; drag it down to close.
• Offline: a bar at the top says when there is no connection. Chats you have opened on the phone still open, and a message you send waits and goes out by itself when you are back online. Signing out removes the offline copies.
• Answer ready: after your first reply you are offered a notification for replies that finish while you are in another app; the app icon also shows a badge. You can turn it off in your browser’s site settings.

Ask more than one AI at once and compare, combine or vote on their answers.

Each extra answer is charged like a normal message. The picker shows a model’s abilities as small icons: privacy, sees images, web search and reasoning.

Press the headset beside the message box to talk to the AI hands-free: you speak, it answers out loud while it is still writing, then listens again — like a phone call.

Illustration: voice mode speaking a reply, with live captions and the Mute and End buttons.

Voice settings (the gear in the voice screen)

• Language you speak — English, বাংলা, हिन्दी, اردو, العربية, Español, Français and more. Replies come back in it.
• Reply voice — Standard uses your device’s voices (free; the most natural ones are listed first and marked ✦). Natural uses the site’s speech models, charged per character like Audio mode. Pick a voice, set the speed and press Test this voice.
• Hands-free — on: it listens again by itself. Off (Tap-to-talk): tap the orb or press Space when it is your turn.
• Interrupt by talking — just start speaking to cut a reply short. Use headphones so it does not hear itself.
• Short spoken replies — asks for brief, conversational answers with no tables, code or links read out. Switch off for full answers.
• Show captions — what you said and what it says, on screen as you talk.

Read-along

Under any reply, Read aloud (the speaker) reads it in your device’s voice one sentence at a time, and the sentence being read is highlighted; the chat scrolls to follow it unless you have just scrolled yourself. Code blocks are skipped. Listen in a natural voice (More → Translate & listen) highlights along with the recording too — it follows by length, so it can run slightly ahead or behind. Turn the highlight off in Settings → Preferences → Appearance → Highlight the sentence being read aloud.

Practice

Switch to Practice, pick an activity, fill in the topic and press Start. The AI leads: one question at a time, then it waits for your spoken answer, marks it ✅ or ❌ with a short explanation, and moves on. A progress bar shows the question number and your score.

• Quiz me — any topic (or “what we discussed in this chat”), 5–20 questions, easy, medium or hard.
• Flashcards — it says the term and you give the meaning, or the other way round; missed cards come back.
• Interview practice — a friendly, standard or tough interviewer for the job you name, with one tip after each answer.
• Language practice — it talks only in the language you are learning, at your level, listens in that language, and corrects you gently.
• Role-play — rehearse a hard conversation; it stays in character and coaches you as you go.
• Spelling & pronunciation — spell the word it says (letters or the word both count), or repeat a phrase after it.
• While you practise: Repeat (hear the question again, free), Hint, Skip and Finish. At the end you get a final score, what you missed with the right answers, and two tips — also kept in the chat. Practise again runs the same set-up.

Record mode

• Switch to Record, choose what you want back — Meeting notes (summary, decisions, action items, follow-up email), Lecture notes (key ideas, terms, a quiz), a tidy Voice memo, or Just the transcript — and press Start recording.
• A live transcript and timer show while it records. Pause and Resume as needed; Stop and summarise sends it to the chat, Discard throws it away.

• The microphone button (in the + sheet on phones) still just types what you say into the message box.

Extra tools sit under every reply (the ⋯ More menu) and in the chat menu at the top.

• Translate one reply, the whole chat, or a document you upload, into any language. Formatting, tables and code stay exactly as they were. You see the cost before it runs.
• Listen reads a reply aloud; Podcast turns it into a two-voice script (short, medium or long) and the audio.
• Versions: edit a question you already sent and the earlier answer is kept. Use ‹ 2/3 › under the message to switch between them; the chat continues from the version you pick.
• Repeat on a schedule: chat menu → Repeat this chat. The same prompt runs every day, weekdays or week at your time (in your timezone), optionally with fresh web results. A bar in the chat lets you pause, resume or stop it.
• Quick image edits: on a generated image, choose Remove background, Replace background, Erase, Extend or Upscale.

Attach a file and ask about it — the answer comes from the file itself.

• Spreadsheets and CSV: the AI works out totals, filters, groupings and comparisons from the real numbers and can draw a chart or give you a cleaned-up file to download. Suggestion chips appear when you attach one.
• PDFs: answers quote the page they came from (“page 12”). Long PDFs are read in full, and picture-only pages are read as images.
• Fill a PDF form: Tools → Fill a PDF form, or Fill this form on an attached PDF. It finds every field — on flat PDFs with no form fields it finds the blanks on the page — suggests answers from what you tell it, and highlights its guesses. Check each one, move or add boxes, then download or save it to Files (optionally locked). Text in any script is supported.
• Meeting notes: More → Meeting notes, then upload the recording. You get a transcript with timestamps and speaker labels; name each voice from a sample of what they said, and the notes list decisions, action items with owners, and open questions.

A knowledge base is a collection the AI answers from (Settings → Knowledge, or Knowledge in the rail).

• Add sources: upload files, add web page addresses, pick Google Drive files (or a whole folder of up to 50) and Notion pages.
• Use it: attach a knowledge base to a chat, a project or a bot. Answers quote the best-matching passages and say where they came from.
• Search by meaning: passages are found by what they mean, not only by matching words. New documents are indexed as you add them; press Make searchable by meaning once for older ones (the cost is shown).
• Keep it fresh: set a base to re-read its web pages, Drive and Notion documents every day or week.

Switch to Draw mode with the image/draw button beside the message box, then describe what you want to create.

Style & look

The Style button (Alt+L) opens one panel with everything that shapes how the picture looks:

• Templates — Logo, App icon, Sticker, Product shot, Thumbnail, Poster, Social post, Wallpaper, Avatar. A template also picks a fitting shape when the model offers it.
• Style, Lighting, Camera, Mood & colour, Composition — pick one of each; click a choice again to clear it.
• Text in the image — words the picture should contain, spelled exactly.
• Reference image strength — Loose, Balanced or Close, for when an image is attached.
• Generate — lock the seed, make the same idea in 3 shapes at once, or write several prompts, one per line, and get one image for each.
• Saved prompts — star a prompt to keep it with its look; your recent image prompts are listed too.
• Brand kit — your colours and style, added when “Use my brand kit” is ticked.

Everything you pick shows as a small tag under the bar; the ✕ on a tag removes it.

Reference images

• Attach with the paperclip button, paste, drag a picture onto the message box, name an uploaded file with @, or choose Use as reference from any generated image’s menu.
• If the selected model cannot use a reference, attaching one switches to a model that can and tells you which.
• A reference is used for one send and then cleared, so the next prompt starts fresh. (Try-On keeps the person photo for the next garment.)
• Click a thumbnail to see it full size; the ✕ removes it.

Working with results

• Edit — describe a change, and optionally paint the area to change.
• From the ··· menu: 2 more like this, 2 close variations, 2 different takes, keep this look (same seed), upscale, remove background, expand to landscape or portrait, compare with the image before, use as a video start image, and download as PNG, JPG or WebP.
• Keys in Draw mode: Alt+L Style & look, Alt+E enhance the prompt, Alt+S next shape, Alt+1 / 2 / 4 number of images.

Make a series

Options → Make a series turns one idea into a whole project of pictures — 30 Instagram posts, 12 picture-book illustrations, a product catalogue, a sticker pack. Every image is made with your current Draw settings: the same model, shape, size, quality, style & look, brand kit and locked seed, so the series matches. Each picture lands in its own chat with a short caption; see Projects & Autopilot.

• Images count against your daily image limit; if it runs out, the rest wait and continue the next day.

Limits to know

Switch to Video mode with the video button beside the message box (or type /video), then describe the video you want.

Direction & look

The Direction button (Alt+L) opens one panel:

• Templates — Product ad, Social hook, Talking head, Food close-up, Property tour, Logo reveal, Nature b-roll, Explainer, Music visualiser. Each sets the format, length, camera and look in one tap.
• Camera, Look, Motion, Lighting & colour grade, Shot — pick one of each; click again to clear.
• Avoid — things you do not want (text, logos, extra people…).
• Sound & dialogue — describe the background sound and write spoken lines, for models that make sound (used when sound is on).
• Generate — lock the seed; make it in 3 formats (Reel 9:16, YouTube 16:9, Square 1:1); a shot list with one video per line; or make the first frame as an image first, approve it, then use it as the start image.
• Saved prompts — favourites (with their direction) and your recent video prompts.
• When it is ready — a browser notification and/or a short chime when a video finishes.

On a finished video

• Download, Regenerate, Extend (Veo 3.1), Make final (after a Draft), Loop and playback speed.
• From the ··· menu: continue from the last frame; save the paused frame as a PNG; use the paused frame as the start or end image, or in Draw; another take; a different take; compare with the video before (side by side, played together); copy the prompt; share.
• Keys in Video mode: Alt+L Direction, Alt+E enhance, Alt+F next format, Alt+D next length.

Make a series

Options → Make a series plans a set of scenes or episodes — a 6-scene product ad, a 10-part tips series, a property tour — and makes each one with your current Video settings: model, length, resolution, shape, sound and Direction & look. Each video lands in its own chat with a caption.

• Videos are made one after another within your plan’s “videos at once” limit: a scene that cannot start yet waits and starts a few minutes later, instead of being skipped.
• The site caps how many videos one series may hold (6 by default) — videos are the most expensive thing to make, so check the estimate before you build.

Limits to know

Switch to Audio mode with the audio button beside the message box (or type /audio), then type the text you want turned into speech.

Voice & delivery

The Voice & delivery button (Alt+L) opens one panel:

• Templates — Podcast intro, Voicemail greeting, Product ad (30s), Explainer, Audiobook chapter, Meditation, Announcement. Each adds a starter script and a fitting delivery; replace the [brackets] with your own words.
• Delivery — Narrator, Newsreader, Friendly, Audiobook, Advert, Calm, Excited, Whisper, Storyteller, Support.
• Pace & pauses — slow and clear or fast, and a short pause wherever there is a blank line.
• Emotion & sounds — buttons that insert the model’s own tags, such as (happy), [whisper] or <laugh>, where your cursor is.
• Accent & language — US, UK, Indian or Australian English, Spanish, French, German, Hindi, Bengali, Arabic, Japanese. Write the script in that language: this sets the accent, it does not translate.
• Pronunciation — your own list, one per line, like SQL → sequel. Saved to your account and applied to every script while it is ticked (file names after @ are left alone).
• Script tools — Fit to 30 / 60 / 90 seconds rewrites the script with AI to that length; Clean for speech removes markdown, links, emoji and code; Undo puts your text back.
• Compare voices — pick up to 3 voices and hear their samples one after another, then use the one you like.
• Saved scripts — favourites (with their delivery and voice) and your recent scripts.

Dialogue

• Up to 4 speakers, each with a name and a voice. Write the script as Name: line, one line each.
• Detect the speakers in my text turns a pasted conversation or interview into that format with AI and fills in the names.

On a finished clip

• Download, Regenerate and playback speed on the card.
• From the ··· menu: copy the text, share, loop, use as a video voice-over, same text with a different voice, a transcript with times (click a line to jump there), download as WAV, and ask about this text in chat.
• Keys in Audio mode: Alt+L Voice & delivery, Alt+P preview the voice, Alt+V next voice, Alt+S next speed.

Make a series

Options → Make a series writes and records a set of episodes — a podcast season, audiobook chapters, daily meditations, language lessons — each as a full spoken script read with your current Audio settings: model, voice or speakers, speed, delivery, accent and (when ticked) your pronunciation list. Long scripts are split and joined automatically.

Limits to know

Canvas opens a reply as a document beside the chat, so you can shape it with the AI and keep it. Open it from Canvas under any reply, from Tools, by typing /canvas, or with Ctrl+Shift+E.

• Share: Word, PDF, Markdown, web page or text; copy as rich text for Gmail or Docs; send to Google Docs or Notion; publish as a page; share with a team.
• Back to the chat: Insert in chat adds the document as a new message; Replace the reply swaps the original answer.
• Templates: blog post, email, report, proposal, cover letter, resume, meeting notes, README and SOP.

When you ask for something you can use — a calculator, a chart, a dashboard, a landing page, a quiz, a small game — it opens as a working Artifact beside the chat.

• It runs in a safe preview: try it, then ask for changes in plain words and the new version replaces it.
• Download it as a file, or Share it: anyone with the link can use it (you can turn the link off).
• Pictures from the chat can be used inside an Artifact.

Turn any chat or topic into study material (Tools → Study, or type /study).

• Audio overview — a two-voice conversation about the topic.
• Mind map, flashcards, quiz and slides made from what you are reading.
• YouTube summary — paste a link, or type /youtube and a link, to summarise a video from its captions.
• Publish as a page — share a result as a public web page you can unpublish later.

Ask for something to happen and the AI sets it up instead of telling you how. Say “remind me tomorrow at 10 about the ABC meeting” and a card appears in the reply with the time it worked out. Press Confirm and it is set. Ordinary questions are still answered as ordinary answers — nothing changes about normal chat.

Illustration: asking for a reminder in chat. The card shows what will happen and when; the green line is what it becomes after you press Confirm.

Try these first

What it can set up for you

• Reminders — one-off or repeating, delivered as a phone notification, an email, a chat message, or all three. Free; they never touch your credits.
• Calendar entries — “I have a meeting with ABC on Thursday at 11.” Anything already booked at that time is flagged on the card, and it can set a reminder a few minutes before.
• Background tasks — “every Monday check these pricing pages and email me a comparison.” The AI plans the steps, does the work and delivers the result, with the tab closed.
• Meeting briefs — “brief me before the ABC meeting.” Shortly before it starts you get a short prep note: what it is about, anything recent about the company or people, and what to be ready to say.
• Page watching — “tell me when their changelog changes.” Checked on a schedule; you only hear about it when something actually moved.
• Email — to yourself now or at a set time, or to one of your saved contacts with the unsubscribe link and compliance footer already handled.
• Contacts, workflows, modules and deep research — anything you have set up elsewhere can be started by asking for it.
• Questions about your own account — “what is on tomorrow?”, “how many credits do I have left?”, “is anything waiting for me?”

Illustration: the Assistant page. Reminders, meetings and scheduled tasks in one list, grouped by day, with anything overdue marked.

How confirming works

Quick ways in

Asking in a sentence always works, but two shortcuts skip the AI entirely — they are instant and cost nothing:

• /remind in the message box. Type the whole thing in one line — /remind call the bank tomorrow at 10 — and a small dialog opens with the time already worked out. Change anything you like, choose notification, email or chat, and set it.
• /agenda shows what is coming up in the next week, with a Done button beside each reminder.

Illustration: /remind reads the time out of the sentence — “tomorrow at 10” becomes a real date you can correct before saving.

• In the Chrome extension, type @remind to be reminded about the page you are reading — pick In 1 hour, This evening, Tomorrow morning or Next week, and the page title and link are kept with the reminder. @agenda lists what is coming up without leaving the panel.

Illustration: the browser side panel. @remind keeps the page you are reading for later, with its title and link attached.

Before a meeting

Ask to be briefed and the assistant schedules a short piece of research that runs shortly before the meeting starts — by default half an hour. It arrives as a notification and an email: what the meeting is about, anything recent it can find about the company or people involved with links, and a few things to be ready to say. It runs the AI, so it spends credits from your balance, and it says so on the card before you agree. Cancel the meeting and the briefing is cancelled with it.

Set your timezone first

“Tomorrow at 10” only means something if the app knows where you are. Open Assistant → Settings and check the timezone; everything the assistant schedules uses it. If your device disagrees with the saved setting, the page offers to switch.

Quiet hours, the daily brief and your calendar

• Quiet hours hold notifications overnight — they wait in your tray instead of buzzing. Emails you specifically asked for still go out on time.
• Daily brief is an optional email each morning: what is on today, anything overdue, and anything waiting for your confirmation.
• Calendar subscription gives you a private link to add to Google, Apple or Outlook Calendar, so assistant events and reminders appear beside everything else. Rotate the link any time if you share it by accident.

What it costs: reminders, calendar entries and organising things are free. Only work that runs the AI — a scheduled task, deep research, a module — uses credits, and the card tells you the budget before you agree to it.

Safety: only your own message can trigger an action. Text inside a web page, a PDF or a file you uploaded can never make the assistant schedule, send or spend anything by itself, and nothing that emails another person is ever sent without you seeing the draft first.

Describe something you want to keep on top of — “help me stay on top of my health”, “I have an exam in six weeks”, “my small-business admin” — and the AI drafts the reminders, calendar entries, scheduled AI tasks and page watches for it. You check every one before anything is created.

Examples: medication and water reminders · exam study blocks with a Friday mock test · invoices, VAT and a competitor price watch · a moving-house countdown · a morning routine · plant watering and chores.

Built-in commands

Type them at the start of a message, or pick them from the / menu:

Your bots

A bot is a helper with a job — a standup writer, an inbox triager, a code reviewer, a language tutor (see them all). Open them from the account menu (🤖 Bots, which shows how many you have on), Tools → Bots in the message box, or the Your bots row on a new chat. Make one in Settings → Bots: install it from the gallery, describe it and let the AI build it, or write it yourself (name, command, instructions, questions it asks first).

• /name — pick it from the / menu; it asks its questions (if any), then answers in the chat.
• @name anywhere in a message — several at once: “@editor tighten this, then @marketer write three headlines” answers as each, in turn.
• Keywords — a word or phrase that applies the bot automatically when it appears in your message.
• On a schedule — every day, weekday, week or month at a set time, into its own chat, with an optional notification and email (and web search when it needs today’s information).
• Webhook — a private URL; anything POSTed to it runs the bot with that text as its input.
• Chat — talk to a bot in a chat of its own: every message in that chat goes to the bot, with its conversation starters as one-tap buttons and its own model if it has one.
• Its own model, conversation starters (up to four) and a memory: a bot can remember facts and preferences between runs — it adds to them by itself, and you can read, edit or clear them in its editor. When it runs, it also sees its last runs, so “what changed since last week” and running logs work.
• Hand-off — after it runs, pass the result to another of your bots (a researcher, then a writer, then a translator; up to three in a row).
• Try it in the editor before saving; Copy a bot; Export it as a file and Import one someone sent you (it opens in the editor so you can read it first; memory, webhooks and knowledge are never exported).
• A bot that fails three times in a row switches itself off and tells you. Keywords match whole words only. What a bot costs in chat counts towards its monthly budget.
• Run now, its results chat and its run log are on the Bots page; chats a bot writes are marked 🤖 Bot in the sidebar.
• Typing @ offers your bots under Helpers; the / menu groups Commands, Your bots, Skills and Shortcuts. A reply a bot answered shows “🤖 Answered by …”.
• Account → Profile shows your bots at a glance and your plan’s bot limits; Account → Credits lists what each bot ran and cost this month. Bot notifications can be switched off on the Bots page.

In messaging apps

In a linked Telegram, WhatsApp, Slack, Discord, Teams or other channel: /remind call the bank tomorrow at 10, /agenda, /bots and /bot name your text — the bot’s answer comes straight back in the chat (and is kept in its chat on the website).

• /menu — everything you have, numbered: shortcuts, bots, modules, tasks, recipes, skills, projects and knowledge bases. Reply with a number or tap a button (Telegram, WhatsApp, Messenger, Instagram, Discord). Call anything by name too: /standup your text. On Discord, /run suggests your things as you type; Telegram adds them to its / menu.
• Questions one at a time: a bot's questions, a module's inputs, a shortcut's variables and a task's fields are asked step by step, with buttons for choices; /skip for optional ones, /cancel to stop.
• /skill name (on/off), /project name (its files and memory apply; /project off), /kb name (answer from a knowledge base; /kb off) for that conversation; /recipe name; /routine pause|resume name; /task pause|resume|last name; /actions what you want (your connected apps).
• /newbot description (drafted by AI, then Save), /newtask description, /newshortcut name: text.
• Confirm first: anything that sends or changes something — an email, an event, a task — shows Confirm / Cancel (or reply yes / no). Personal things (tasks, recipes, projects, knowledge, actions) work only in a private chat with the bot, never in a group. Deep Research is not offered in messaging apps.
• /use name — talk to one of your bots in that app: every message goes to it until /use off; replies are marked with the bot’s name.
• /image a description, /search a question (when the channel allows web search), /translate french text, /summary of the conversation.
• /remember a fact and /memory to see what it remembers; /chats lists your recent chats from that app and /open 2 carries one on.
• On the website too: every conversation from Telegram, WhatsApp, Messenger, Instagram, Discord, Slack, Teams or email is a chat here, with a header showing the app, where it is (direct messages, a group or a channel), your account there, when the last message came and, on WhatsApp, Messenger and Instagram, how long replies are still possible. Open in … jumps to it in the app; new messages appear while the chat is open; the sidebar shows unread counts. Search discord (or any app, or messaging apps) to list those chats.
• Reply from the website (off unless you switch it on, per chat, under ⋯ in the header): what you ask and get answered here is also posted in the app — private chats only unless the site allows groups, and on WhatsApp, Messenger and Instagram only within 24 hours of their last message.
• In groups the AI knows which person asked and keeps its reply short. On Slack, start commands with ! (for example !image a fox); Telegram and Discord show the commands in their menu.

A chat answers you now. A task goes away and does the work — searching, reading pages, writing, building a document — and comes back with the result. It runs on the server, so you can close the tab, and it can run again every week without you. The Tasks page lists every template, the three planning modes and the five delivery channels.

Making one

Or skip the form entirely: say it in chat. “Every Monday check these pricing pages and email me a comparison” builds the same task and shows you a card to confirm — see Assistant & reminders.

Illustration: a task card and the timeline of one run — every step, how long it took and what it cost, with the delivered file at the end.

Worked examples

While it runs, and after

• Each run shows a timeline: every step, how long it took, what it cost, and the output it produced. A failed step can be retried on its own.
• Approvals are yours to set: approve the plan before it starts, approve the result before it is delivered, both, or neither.
• Each run has a credit budget. It stops when the budget is reached rather than spending on quietly.
• Anything a run produced — a PDF, a spreadsheet, an image — is attached to the run and to wherever it was delivered.

Building it your way

• Describe it: press ✨ Describe it and say what you want in a sentence — the AI fills in the goal, steps, schedule, delivery and budget for you to check. From any chat, ⋯ → Turn into a task does the same with that conversation.
• Steps for every job: besides search, reading pages and writing, a step can do deep research, run Python in a code sandbox, read a CSV or Google Sheet, call an API, read a PDF (scans too), look at images, search your project files, documents and notes, use your connected apps (Google, Notion, MCP), hand work to one of your bots or modules, make a video, or post to social media after you approve.
• Logic: a Condition step continues, stops quietly or takes a yes/no branch; For each repeats a step for every item, address or row; Ask me pauses for your answer (on your phone too). Steps can save their output as {{name}}, use their own model, retry on a cheaper model, have a time limit, or run together.
• Fill-in fields such as a company name are asked each time you run it. Test run does everything except deliver. The editor shows what a run and a month at that schedule should cost.
• More ways to start: several times a day, weekdays only, quiet hours or a cron line; a new item in an RSS feed; an email to the task's own address; before or after a meeting; a new project file; /task name in Telegram, WhatsApp or Slack; another task finishing; or another app through its webhook or the API.
• Results your way: a report, table, bullet digest, brief or email, in any language, also as PDF, Word, Excel, PowerPoint or CSV; saved into Notes, a document or a project as a running log; sent to Google Docs, Notion, Slack, Discord, Teams or by text message to clients; only when something changed or mentions a phrase; or gathered into a daily or weekly digest.
• Keeping track: a Dashboard (success rate, spend, failures, what runs next), a Calendar of past and upcoming runs, search across every result, side-by-side comparison of two runs, comments, version history with restore, CSV export, and alerts when a run fails or a budget is nearly used. Set a monthly budget per task or for all of them.
• While it runs: each run shows its stage, the step it is on and that step's prompt, a progress bar with a time estimate, credits used against the budget, and a live activity feed ("Searched…", "Read 5 pages", "Step 3 done"). You can skip a step, edit the steps that have not started, add credits or time, send it a note ("focus on prices only"), or finish now with what it has. Running tasks also show in the tab title, on the Tasks icon in the chat page's rail, and as a badge on the Chrome extension.
• Sharing: share a task with your team (who may see, run or edit it, and who approves), send a template link, or publish it to the Community gallery for others to copy.

Emailing clients

A task can send its result to contacts you keep in Tasks → Contacts, imported from CSV or added one at a time. That side is deliberately careful: you add your business name and postal address first, every message carries them and an unsubscribe link, new senders have their mail reviewed before it goes, daily limits apply, and bounces and spam reports pause sending automatically. You can send through the site or connect your own mail server so it comes from your own domain.

What it costs: a run spends credits like a chat does — more for more steps. The budget you set is the ceiling, and Tasks will not start a run if your balance is below the site minimum.

A skill is a set of instructions for one kind of job — how you want invoices laid out, what your code reviews must check, the tone your newsletters use. Write it once and the AI follows it whenever a matching request comes up, without you pasting the same briefing again. 20 are built in and install in a click; the Skills page lists them all with the request each one answers.

How the AI picks one

• Only a one-line description of each skill is sent with your message. The full instructions are loaded only when one is actually used, so having a shelf full of them costs you almost nothing per message.
• Type /skill-name to force a particular one.
• Turn on Always on a skill's card and it applies to every message.
• Otherwise the AI chooses by itself from those descriptions — which is why the description should say when to use it, not just what it is.

Getting them

Illustration: the tester. One message, two answers — with your skill and without it — plus the tokens, credits and time each took.

Two ways to start

• Ready-made — Skills → Library has 20 working skills across Writing, Marketing, Business, Development, Analysis and Productivity: email replies, meeting notes, invoices, code review, SQL, proofreading, plain English, bug reports, test cases, spreadsheet formulas, interview guides, decision briefs, study notes and more. Search it, install one, and it works in every chat straight away. You can edit it afterwards like your own.
• Write with AI — describe the job (and paste an example if you have one); the AI drafts the name, the “when to use” description, the instructions and any fill-in fields. You check it before saving.
• Save a chat as a skill — in a chat’s ⋯ menu. And when you send near-identical instructions a third time, the chat offers to save them as a skill.
• Description score — the editor rates how likely the description is to get the skill picked, with tips, and ✨ Improve with AI rewrites it.
• Fill-in fields — a skill can ask for details that differ per person (company name, tone). Fill them once under the skill’s Your details tab.
• Per chat — the Skills button in the message box sets each skill to Auto, Always or Off for that chat only. Replies show which skills shaped them, with 👍 / 👎.
• Team skills — team owners and admins can make a skill for the whole team (choose the team under Available in).
• Tests and stats — save a few test messages with what a good answer must include; Run all compares with and without the skill and marks pass or fail. The Stats tab shows loads, tokens and ratings, and My usage covers all your skills.
• Community — share a skill from its ⋯ menu; once approved, others find it in Community, install a copy and rate it. Packs install several skills in one go.
• Everywhere — skills also apply in bots, modules, background tasks and messaging apps. Write /skill-name in a bot’s or module’s instructions to always use that skill. When a model can run code, a skill’s scripts run in the provider’s sandbox.
• From a shape — + New skill now asks what kind of skill you want first: a house format (invoices, reports), a review checklist, a reply in your voice, a detail extractor for messy text, a document with fixed sections, a rewrite to your tone, or answers from your own rules. Pick the nearest one and the name, the description and a first draft of the instructions are already written — you edit rather than start at a blank page. An empty skill is still one click away.

A skill that works, in three lines

The description is the part that matters most: it is the only text the AI sees when deciding whether a skill applies. “Use when…” belongs in it.

Keeping them current

A skill imported from GitHub can check for updates. You see the changes line by line before applying them, and the version you were using is kept so you can roll back. Skills can also be shared to the site library for other people to use, once an administrator approves.

Skills change how the AI writes; they never grant it new powers. A skill cannot give itself tools, reach your files, or override the rules the site runs under — and a task can be told to follow specific skills too.

Modules are guided, fill-in-the-blank AI templates that turn a task like “write a Facebook post” into a short list of questions instead of a blank prompt box. Answer them and the AI does the rest. 265 are included, across 27 categories, and you can build your own.

• Each module can restrict which AI model it runs on — if it needs one your plan doesn't include, you'll see that clearly before submitting rather than after.
• Image modules use the same Draw pipeline and credit cost as generating an image manually.
• A module that's still mid-way through its questions can be cancelled at any point without sending anything.

What is in the library

Each category is searchable, and the library page has a filter chip per category:

• Daily & weekly routines (10) · Planning & focus (9) · Meetings & prep (9) · Inbox & messages (8) · Watch & alerts (8) · Team & management (8) · Freelance & clients (8)
• Writing & editing (12) · Marketing & ads (20) · Social & short-form (14) · SEO & content (12) · Sales & CRM (8) · Customer support (8) · Business & operations (16)
• E-commerce (11) · Finance & analysis (9) · Legal & compliance (9) · Career & HR (10) · Teaching & study (11) · Learning routines (6) · Research & data (9)
• Developer & technical (10) · Creative & media (9) · Property (7) · Bangladesh & regional SME (8) · Home & family (8) · Personal & everyday (8)

Modules for work that repeats

Most of the library is for one-off jobs. The routine categories — daily and weekly routines, planning, meetings, inbox, watch and alerts, team, freelance, home and learning — are written for repetition instead: each one reports what changed since the last run rather than restating everything, and says so plainly when nothing did. A morning brief, a standup note, a competitor watch, an inbox triage, a weekly review.

• Any module can be given a schedule — hourly to monthly, in your timezone — and then runs in the background and delivers the result without you opening the app.
• Every firing re-checks your plan and your daily run limit, so a lapsed entitlement stops the schedule rather than quietly spending.
• A handful are multi-step packs: several runs chained together, each reading the one before it — research, then draft, then a finished document. 17 of them, and your plan sets how many steps you may chain.

A project is a workspace for one job — a client, a course, a product launch. Chats inside it share the same instructions, notes, memory and files, so every new conversation starts already knowing the background instead of you re-explaining it each time.

• Project memory is separate from your account memory. Facts learned inside a project stay in it, so nothing from a client project leaks into an unrelated chat.
• Each project chooses how the two combine: merge both, use only its own memory, or use only your account memory.
• Pinned notes hold the current state of the work — decisions made, things still open — and are sent with every message in the project.
• Its own model. A project can pick the model its chats start on, so research work opens on a strong reasoning model while routine drafting opens on a cheaper one. Tick Always use this model in this project and opening any chat there switches the composer to it, not just new ones — you can still change it for a single message.
• Knowledge files are searched by relevance, so only the passages matching your question are included rather than whole documents.
• The sidebar stays where you put it. Picking a project filters your chat list to that project; opening a project chat from All chats does not — the chip above the composer tells you which workspace the current chat belongs to, and Close leaves the project without touching the list.
• Sharing and collaboration. Send a read-only link to a single conversation, or invite someone in as a collaborator to work alongside you in a chat or a whole project. In a chat, each person pays for the AI replies to their own messages; remove access when the work is done.
• Group chats. Tools → Group chat turns a chat into a group: share an invite link (anyone signed in who opens it joins, up to 20 people) or invite by email. Everyone sees who wrote each message, and new messages arrive by themselves. You choose when the AI answers — every message, or only when someone writes @AI (or @ one of their bots) — and the AI knows who is in the chat and who asked. Away for a while? Catch me up summarises what you missed (a little AI, billed to you). Personal memory is never used or added to in a group chat. Leave a group any time from Manage.
• Reactions, Reply and names in a group chat. Under any message, press the smiley (on a phone, also in the press-and-hold menu) to react with 👍 ❤️ 😂 😮 🎉 or 👀 — everyone sees them, and pressing yours again takes it back. Reply on someone else’s message quotes it in your answer and puts their @name in front. Type @ to pick a person in the chat (or @AI); naming someone sends them a notification that opens the chat.

Autopilot: a project that builds and runs itself

Describe a goal once — “30-day home workout for a beginner, 20 minutes a day, bad knee” or “help me run my cafe” — and the AI drafts the whole project for you to check before anything is made or charged.

• Chats are a sequence you work through — Day 1–30, Lesson 1–12, Chapter 1–15, Post 1–20. Each is a normal chat in the project, already written, with its picture, narration or video.
• Agents are standing roles — Coach, Inventory manager, Marketing, Editor — that work on their own schedule (every day, weekdays or weekly at a set time). Each writes its update into its own chat and a one-line summary into the Project hub chat. You can talk to an agent in its chat any time; it answers as that role.
• An agent allowed to adjust upcoming chats (a coach or an editor) can add notes to chats that have not been made yet — for example making next week’s workouts easier.
• Projects you made by hand can use Autopilot too: Autopilot → Add chats with AI.
• From the chat: Tools → Project, typing /project and your goal, or the “✨ Build this as a project” suggestion that appears when you describe a series (“a 12-week course”, “30 posts”). In Draw, Video and Audio mode, Options → Make a series builds a series with that mode’s current settings. The chat stays open: you land in the new project’s hub and a small bar shows progress.
• Waiting: a chat whose picture, narration or video is held back by a limit shows Waiting for limit and finishes by itself — a few minutes later for a busy video slot, the next day for a daily limit.

Examples: a 30-day workout with a weekly coach · running a cafe (menu, stock, marketing, rota, reviews, money) · learning a language in 12 weeks with a daily tutor · a 6-week product launch · a month of social posts with a trends agent · writing a book with an editor · a 90-day study plan with weekly mock tests · planning a wedding · a job search · running an online shop.

Autopilot limits

Chat with AI Chat Console from Telegram, WhatsApp, Messenger, Instagram, Discord, Slack, Microsoft Teams or email. It is the same account — same models, memory, bots and credits — and every conversation is saved on the website. Messaging apps are part of paid plans, and the site chooses which apps are switched on. More on the Messaging apps page.

Your things, from the app

• /menu lists your shortcuts, bots, modules, tasks, recipes, skills, projects and knowledge bases. Reply with a number or tap a button; WhatsApp shows long menus as a tap-to-pick list.
• Call anything by name: /standup, /bot editor tighten this, /task weekly report. A bot or module that needs details asks for them one at a time; reply skip or cancel.
• /use tutor sends every message to that bot until /use off. /task pause, resume and last manage a task.
• /newbot, /newtask, /newshortcut make one from a sentence.
• /actions what you want uses your connected apps. Anything that sends or changes something comes back as a Confirm card; nothing happens until you press it.
• Discord’s /run suggests your things as you start typing the name.
• Admin settings are never available from a messaging app.

On the website

• App chats show in the sidebar with where they came from (“Discord · #general”) and an unread count.
• Opening one shows a header: the app, direct message, group or channel, the account, when the last message came, and Open in the app. New messages from the app appear while it is open.
• Also send replies there: switch it on in the header and what you ask and get answered on the website is posted in the app too (when the site allows it).
• 24-hour window: WhatsApp, Messenger and Instagram only allow replies within 24 hours of your last message there. The header shows how long is left.
• Under Settings → Connected Apps → Messaging apps each linked app has its own Model, an option to Send task results and reminders here (if the site offers it), and Unlink.

The chat knows what your connected apps can do, and offers the next step. Connect a new app — or any MCP server — and its features show up here by themselves.

While you type

• Suggestions beside your text: as you type, small icons appear right after your last word (on a phone, a 💡 with a count sits next to the voice-chat button — tap it for the list). They show what fits your message — Search the web, Make it a file, Make an image, Deep research, the right action from a connected app (for example Send email · Google) with the email address and date you typed already filled in, Use skill, Use a knowledge base, Set it up every Monday, Watch for it. Click one to use it for this message. Checking costs nothing — no AI is used.
• Pasted content: paste a table and get Chart this; an error log, Find the problem; code, Review this code; an email, Draft a reply; a YouTube or PDF link, Summarise.
• Type # to point the message at one of your apps (#Google, #Slack), a knowledge base, a skill, or — for a new chat — a project. The AI then uses exactly that.
• Before it sends: if the message needs an app you have not connected, a card offers Connect (sign in in a small window and the message sends by itself) or Answer without it. If two of your accounts could do it (say Gmail and Outlook) it asks which. If a connection has stopped working it offers to reconnect.
• Several things at once: “summarise this and email it to Sara” → Do it as steps plans each step from your apps as one card you check and confirm, and can save as a recipe. This uses a little AI, billed to your credits, with a daily limit you set.
• Also: Tab finishes a message you have sent before; Make this clearer rewrites a short, vague request (a little AI); a request you make often gets Save as a shortcut; web search and file making switch on by themselves when your message plainly needs them.
• Under the reply: one line shows what the answer used (web search, an app, a skill, your knowledge); a reply with a table gets Download CSV and Excel file.
• Right-click (or long-press) a chip to stop suggesting it. Everything can be switched on or off in Settings → Smart actions → Smarter chat.

After a reply

• Next steps under the reply: right under the newest answer and its sources, a Next steps line offers up to three suggestions (on a phone, one line you swipe) such as Draft reply in Outlook, Add to Google Calendar, Make a PowerPoint or Open GitHub issue, with the details filled in from the chat. Click one to check and change them first.
• Safe by design: things that only look something up run at once and the result appears in the chat. Anything that sends, posts or changes something shows a card you confirm. Nothing is ever deleted by a suggestion. Values taken from pasted emails or web pages are highlighted so you check them.
• Several steps at once: e.g. make the slides → save them to OneDrive → email the link — one card, confirmed once, then it runs in order. Save it as a recipe to run again by hand or on a schedule.
• Not connected yet? The chip says connect Outlook; sign in in the small window and the action carries on.
• Everything at hand: Tools → Actions (or Ctrl+K → Actions) lists and searches every action you have.
• In conversation: just ask — “draft that as an Outlook email”, “save the deck to my Drive” — and the AI finds the right action itself.
• Settings → Smart actions: turn suggestions on or off, switch off any action, let it check chosen apps every few hours for things to do, and manage recipes. It learns from what you use and ignore.

Let the AI use your own tools during a chat (Settings → Connectors). It searches and reads them freely; anything that writes — an email, an event, a post, a document — first appears as a card, and happens only when you press Confirm. It never deletes anything.

• Google: search and read Drive, Gmail and your calendar; draft or send Gmail, add calendar events (with invites) and create Google Docs. Connected before these were added? Reconnect Google once to allow writing.
• Microsoft 365: search and read Outlook mail, your calendar and OneDrive; draft or send email and add events.
• Notion: search and read pages you share with the app; create pages.
• GitHub, Slack, Airtable: paste a personal token. Search issues and read files, read channels, read records; open issues and comment, post messages, add records.
• Dropbox: search and read files; save new text files.
• MCP servers: add any MCP server address to give the AI its tools.
• Pages that are built by JavaScript or that block simple requests can be read in a real browser when the site has it switched on.

Put any of your bots on your own website, so visitors can chat with it.

• Open the bot (Bots → edit) and choose Put this bot on your website.
• List your website addresses — the chat only appears on those sites and their subdomains. Pick a name, colour and greeting.
• Copy the one line of code and paste it before </body> on your site. A chat button appears in the corner.
• The bot answers from its instructions and knowledge bases. Every reply is charged to your credits, within the daily limits you set (replies and credits), with extra limits per visitor.
• See conversations in the same window lists every visitor chat.

Connect the service to thousands of other apps (Settings → Connected Apps → Automation).

• Create a token and paste it into Zapier, Make or n8n as a Bearer token. The page shows the address to use.
• Actions: start one of your tasks, run a bot, or ask the AI a question — billed to your credits as usual.
• Triggers: get an instant call when a task finishes, fails, starts or is waiting for you, or when a bot finishes.
• Every call is signed so the receiving app can check it came from us. Remove a token any time to cut access.

In a task with fixed steps (and in Review the plan), switch between List and Flow above the steps.

• Flow draws the steps from Start to Deliver. Steps that run together sit side by side; a Condition’s Yes and No steps get a lane each.
• Drag a box onto another to move the step; click a box to edit it; add steps from the row of buttons underneath.
• Your choice of view is remembered. While a task runs, the write-up appears word by word as it is written.

Work with other people (Settings → Teams): create a team, invite people by their account email, and they join when they accept.

• Share a project with everyone at once, or start a team chat everyone can write in.
• Share a Canvas document with a team so members can open and edit it.
• Owners and admins see what each member used this month. Each person pays for their own messages.
• Community bots: offer a bot to everyone from the Bots page; once approved, others install their own copy.

The extension puts the same chat in a side panel next to whatever you are reading, so you can ask about a page without copying it into another tab.

• Model picking, web search, reasoning and document generation all work the same as on the web app.
• Images and video can be generated from the panel too.
• Projects and the per-chat memory menu (Memory on, Memory off, Independent) are available there as well, so a chat started in the panel behaves exactly like one started on the site.
• Type @ to list your bots; @youtube on a YouTube tab attaches the video’s captions to summarise.
• Every reply has Save note and Save as document (opens it in Canvas on the site), and the ghost icon starts a temporary chat.

Sometimes a model is busy, down for maintenance, or retired by its provider. Normally you would see an error and pick another model yourself. Switch on automatic fallback and the app does that for you: your message goes to the closest match instead, and the reply says so.

It is off by default. Answering with a different model than the one you chose should be your decision.

Illustration: the badge still shows the model you picked; the dashed note underneath shows who actually answered. Your chat stays on your model — the next message tries it first again.

How the replacement is chosen

• It can do what your message needs. An attached image needs a model that can see; a message that uses tools, the assistant or skills needs one that can call tools; web search stays with the same provider so your search still runs.
• Same or similar price. Never more than 1.5× what your model costs, and cheaper is always fine. A free model only falls back to another free one.
• Similar size. A small, fast model is replaced by another small, fast one, not by the most expensive model on the list.
• Somewhere that is working. If a whole provider is down, a different provider is preferred. If one model was retired, a newer model from the same family is preferred.
• Your privacy settings hold. With Zero Data Retention on, only ZDR models are considered.
• A limited number of tries — two other models and about 40 seconds by default. After that you see the error rather than waiting longer.

What it never does

• Use your own API key. If your message was running on a key you added under Settings → API Keys, you get the error — the app never quietly moves your request somewhere else, and never onto your key.
• Switch after the reply has started. If a model fails halfway through writing, you keep what arrived.
• Retry a refusal. If a model declines a request on content grounds, the app does not go looking for one that says yes.
• Retry a message that is too long. A prompt over the model’s limit would fail the same way elsewhere — shorten it, or start a new chat.

Worked examples

The service runs on credits. Each message uses a small number of credits depending on the model — cheaper models go much further.

• Plans run on a 30-day cycle counted from the day you subscribe, not from the 1st of the month. Each cycle grants your credit allowance.
• An invoice is issued at the start of every cycle, before you pay it — find it under Account → Billing, with a PDF to download.
• Nothing is auto-charged to a stored card unless you switch on auto top-up for credit packs. You renew each cycle yourself, through whichever payment method you prefer.
• If a cycle goes unpaid there is a short grace window with full access, after which the account moves to the Free plan. Nothing is deleted — chats, projects and any credits already in your balance stay.
• Your plan's monthly credits renew each cycle and are used first. What is left at renewal does not carry over.
• Credit packs are one-time top-ups that last 12 months from the day you buy them. The oldest pack is used first, and Account → Credits shows when the next ones expire.
• Your current balance is shown in the app; top up anytime from Account → Buy Credits.
• Daily fair-use limits keep the service fast for everyone — see the Fair Usage Policy.
• For renewals, invoices, changing plan and what happens when a cycle goes unpaid, see the Billing guide.

See where your credits go and never run out by surprise.

• Your usage (Settings → Usage): credits used over 7, 30 or 90 days or a year, by day and by feature, your top models, the chats that used the most, and how long your balance lasts at this rate.
• Low-balance alerts (Credits & Usage → Low balance & auto top-up): choose a level and get one notification, email and push when you fall below it.
• Auto top-up (optional): buy any credit pack by card with Save this card for auto top-up ticked, then choose the pack, the level and how many times a month at most. It is invoiced like any purchase, you are told each time, and it switches itself off if the card fails twice. Remove the card any time.

Share your link from Settings → Referrals (WhatsApp, X, LinkedIn, Facebook, email or your phone’s share menu).

• When a friend signs up through it and confirms their email, you both get credits.
• You also earn a share of the credits on their first purchases.
• The Referrals page lists who joined, how far each got (signed up, confirmed, bought) and what you earned.

Your prompts are sent to the AI provider you pick, used to generate a reply, and returned to you. Saved chats stay in your account until you delete them.

• Delete any chat — or your whole history — at any time.
• Export your data in several formats before deleting.
• API keys you add are encrypted and never displayed in full.