Dictation for Slack
Most of a working day in Slack is answering people. Somebody asks why the deploy is red, somebody asks whether the migration can wait, somebody needs the reason behind a decision written out so the rest of the channel can stop guessing. It is the part of the job with the most words in it, it is almost all prose, and it is far quicker said than typed — especially the long thread reply you keep putting off because it will take four minutes to write.
Deixis is voice input for exactly that. Hold Ctrl+Win, say it, let go, and clean punctuated text
lands in the message box you were already looking at, with your clipboard handed back the way you
left it. Nothing to switch to, no transcript to copy over, no account to make.
And Deixis is not a Slack app, a bot or an extension for anything: it types wherever your cursor already is. So the one key dictates your Slack reply, your commit message, your email, your ticket and your browser's address bar — one voice input for the whole machine, not one per app. It runs on Windows, Mac, iPhone and Android, with speech recognition on the device every time, and the audio never leaves it.
Why Deixis
What it gives you in Slack, and what it will not do
Why Deixis in Slack
- A reply is punctuated like a reply, not like an email. Deixis knows which app and which field it is typing into, because the name of the app and the name of the control ride along with your words — never the window title, and never anything on your screen. "yeah that's fine ship it after the migration lands" stays one line in a channel. The same voice in a thread, explaining why the poller double-counts, comes out as three sentences. That difference is the whole point.
- The long thread reply stops being a chore. Slack's own guidance is to take a detailed discussion into a thread so it does not clutter the channel — which is a polite way of saying somebody has to write four paragraphs. Talking them takes a fraction of that, and you get to think out loud instead of composing.
- Talk and point, for the messages that carry a screenshot. Double-tap
Ctrl+Winand Deixis keeps listening while you drag a region of the screen, pick an element or grab a selection. What you said and what you pointed at come out as one note, pasted at the caret. Most Slack messages that matter are a picture and an explanation; this is both of them in one pass. - Grab the region, mark it up, hide the customer name.
Ctrl+Win+Sfreezes the screen and lets you drag a region — quick-save it, or open the editor for arrows, boxes, a highlighter, numbered step badges and a pixelate tool to redact whatever should not be in a channel. Copy puts a real image on the clipboard, ready to drop straight into the message box. - Record it instead of writing a wall of text.
Ctrl+Win+Vrecords the screen you are on with your narration, writes the transcript up on your own machine, and gives you a link anyone can open without an account of their own. A small bar shows it is rolling, with pause and stop, and you can drag it out of the way. When the answer is "watch me do it", paste the link. - Meetings, recorded and transcribed on your machine. A call is a call:
Ctrl+Win+Mrecords your microphone and, where it can, the system output, transcribes both on this machine, and writes a readable transcript into a folder you own. It tells the remote voices apart and remembers the ones you name, and the voiceprint that does it stays on the machine, is never uploaded, and can be deleted per person. - Your own buttons, on the same palette. Add a button that runs your program with the finished note and types its output back at the caret — file the ticket and get the key back, open the branch, look something up. Or chain steps together as plain settings: read the text off a screenshot, translate it, summarise it, post it. No scripting language, and the Buttons panel has ready-made ones a click adds.
- A signed webhook to your own URL. Every finished note, grab or translation can be posted to an endpoint you own, with an HMAC signature and a timestamp so your receiver can prove it came from you and refuse a replay. Delivery retries in the background and never makes you wait.
- It learns your words. Product names, people's names, the library your team argues about weekly — one key press each, fixed on your own machine before anything is sent anywhere. Five vocabulary packs ship in the box.
- Free, with no account, forever. Dictation, both speech models, the dictionary, screen grabs, screen recording, meetings, your own buttons and webhooks: all free, and none of it behind a sign-in. Everything it makes is an ordinary file in one folder you can back up or delete.
What it will not do
- It does not know your channels, your emoji or your people. Deixis types text and stops. It will not turn "tell the platform channel" into a channel link, will not find the right person to @-mention, and will not pick an emoji for you. Those are still your fingers.
- It does not read your screen. Deixis is dictation, not an agent. It sends the name of the app and the field, never the contents, so it cannot see the message you are replying to and cannot answer it for you. You still have to say what you mean.
- English only. The speech models are English models. Slack's own dictation, in its assistant prompt, says it works with the languages Slack supports and many others while being optimised for English — if you work in another language, that is theirs and not yet ours.
- It will not transcribe a file you hand it. Deixis transcribes what it records: your dictation, your meetings, your screen recordings. Drop a clip somebody else posted into it and nothing happens — for that, Slack's own Generate transcript button on the clip is the tool.
- No Linux build, and no PC older than 2013. The speech engine is compiled against an instruction set that arrived in 2013, so an older processor stops rather than running slowly. Linux is not planned, and Slack does ship a Linux app.
- It is not a keyboard replacement. Code snippets, paths and flags are still faster typed. Deixis is for the prose half of Slack, which happens to be most of it.
What you get
More than dictation. One key, and everything behind it.
Hold the key and press a letter: free mode, a grab, a meeting, a recording, a translation, your own buttons. All of it on your own device, all of it free, none of it behind an account.
Dictation, where your cursor is
Hold a key, speak, release. Clean, punctuated text lands in whatever has focus, on your PC, your Mac and your phone, with your clipboard handed back.
Free mode: talk and point
Narrate for as long as you like while you grab a region, pick a button on screen or take a selection. One note, what you said interleaved with what you pointed at, pasted where you finish and saved as Markdown.
Screen grabs, annotated
Freeze every screen at the keypress, drag a region, add arrows, boxes, text and numbered steps, and pixelate what is private. Saved as a PNG that stays editable.
Screen recording, shared as a link
Record the screen with your narration. The recording writes itself up with a title, chapters and a summary, and from your Deixis account it shares as a link anyone can open without one. The Loom you can stop paying for.
Meetings, no bot, names included
Your microphone and the other side, transcribed on your machine while the call runs. Deixis tells the voices apart and remembers the ones you name, and never uploads a voice.
Your own buttons and workflows
Add a button to the palette: it gets the note on stdin and types back whatever it prints. Send finished notes to any URL you own. Chain steps: read the text off a screenshot, translate it, summarise it, post it.
Field by field
The same voice, punctuated for wherever you are in Slack
Deixis knows which app and which field your words are going into — the name of the app and the name of the control, never the window title and never what is on your screen. A prompt, a message and a commit line therefore come out differently. Here is the message box, the thread and the canvas, kind by kind.
| Where you are | What you say | What Deixis types |
|---|---|---|
| A quick channel replythe message box in a channel | “um yeah that's fine ship it after the migration lands” | Yeah, that's fine — ship it after the migration lands.A one-line reply stays a one-line reply. Deixis knows it is typing into a chat box, so it punctuates the sentence and stops there. |
| A long thread replythe thread reply box, under the message you are answering | “so the reason the poller is double counting is that the feed is read once on start and once on the first tick um we only noticed because the release page showed two entries the fix is to skip the first tick and i'd rather do that than cache the response” | The reason the poller is double-counting is that the feed is read once on start and once on the first tick. We only noticed because the release page showed two entries. The fix is to skip the first tick, and I'd rather do that than cache the response.The same voice, a different field. A thread reply is an explanation, so it comes out as sentences rather than one long line. |
| A direct messagea DM with one person | “can you look at the deploy before you log off no rush if it's late” | Can you look at the deploy before you log off? No rush if it's late.A question mark you did not say, because the sentence is a question. A DM reads like speech and is punctuated like speech. |
| A canvasa channel canvas, or a canvas attached to a huddle | “decisions we're moving the cleanup call off the hot path we're keeping the rules fallback and shad kin stays for the settings panel” | Decisions: we're moving the cleanup call off the hot path, we're keeping the rules fallback, and shadcn stays for the settings panel.A document field gets document punctuation — and your dictionary fixes shadcn on this machine, before anything is sent anywhere. |
| Your statusthe status field on your profile | “heads down on the release until three” | Heads down on the release until threeA short label field gets a capital and no full stop, because a status is a label rather than a sentence. |
Slack facts from downloads, threads, slackbot, clips, huddles, huddle_notes, canvas, files and status, read on 7 September 2026. Everything about Deixis on this page is from its README for the current release.
Worth knowing about Slack
- Slack's own help pages describe desktop apps for Mac, Windows and Linux and mobile apps for iOS and Android — five places your caret can be, and Deixis is a plugin for none of them. It types where the caret is.
- Threads exist so that a longer answer does not clutter the channel: Slack's guidance is to reply in a thread to "discuss a topic in more detail", with a box you can tick to send the reply back to the channel as well. The short reply and the long explanation are two different fields, a click apart.
- Slack has voice of its own, in one specific place: its Slackbot article says you can "tap the microphone icon to dictate your prompt instead", that dictation "works with all Slack supported languages and many others, though it's currently optimized for English", and that "Dictated messages are transcribed, but audio data is not stored". That is the assistant's prompt box, not the channel message box, and the page does not say where the audio is transcribed.
- Audio and video clips are the closest thing Slack has to talking into a channel: up to five minutes, available on all plans, playable by anyone in the channel or DM. They are not transcribed automatically — you press Generate transcript on the clip afterwards — and closed captions are available for all video clips.
- Huddles are calls, not dictation. On the free plan a huddle is two people; paid plans take up to 50, with 25 on video. Live captions are not saved once the huddle ends and are English only, and each huddle carries a canvas for notes.
- Slack's AI huddle notes — the feature that writes the call up and embeds a transcript in the notes canvas — are "available on paid plans", and the transcript does not show up in search results.
- Canvases are "surfaces built into Slack where members can create and share fully-formatted content". Channel and DM canvases are on every plan; standalone canvases are a paid-plan feature.
- A status is a field too. Slack's own page says a status can be set to clear itself with "Remove status after…", and that if you pick no emoji a speech balloon is used — so the only part that is yours to write is the words.
- Most Slack messages that matter carry a picture: you can drag up to ten files into the message box at once, add a message alongside them, and write a description on an image.
Five minutes
Setting it up
About five minutes, and nothing is installed into Slack at all — no app to approve, no bot in your workspace, nothing for an admin to review.
Install and fetch the model
One installer, and nothing else to install alongside it. The first launch shows a single Download button for the speech model — about the size of a short video, fetched once, and checked before it is kept. From then on you are offline unless you choose otherwise.
The gesture
Hold Ctrl+Win, speak, release. That is the product. Double-tap it instead and Deixis keeps
listening until you double-tap again — that is free mode, and it is the one for a thread reply you
have not finished thinking through, because you can grab a region of the screen mid-sentence and it
joins the note.
Hold the chord a moment longer and a small palette appears with a letter per mode: S grabs and
annotates, V records the screen, M records a meeting, T translates what you have selected, and
D puts the word under your cursor into your dictionary. If the chord fights something in your
setup, Settings changes it.
Teach it your words
The biggest single improvement, and it costs nothing in speed. Select a word anywhere — a colleague's
surname in a message, a service name in a dashboard — press Ctrl+Win+D, and it is in your
vocabulary. A name it keeps mishearing can be corrected once and stays corrected. The Dictionary
tab is the same list in one place, and the five packs that ship are a click each.
Decide what may leave, once
Settings → Cleanup is the only privacy decision to make, and its default is the quiet one: rules tidy your words on this machine unless you asked for a transform or dictated a long stretch with no punctuation. Set it to never, or leave the endpoint empty, and dictation never touches the network at all. Point it at a model running on your own machine and the better prose stays local too. Whatever you pick, the audio is not part of it — that never leaves, and there is no setting that makes it.
Your own buttons and webhooks
Worth ten minutes once. A palette button can run a program with your finished note and paste back whatever it prints, so "file a bug about the poller reading the feed twice" becomes a ticket and the ticket's key lands at your caret, ready to drop into the channel. Your script gets the note as JSON, arguments are handed to the process as a list so nothing needs escaping, and the first run shows you the exact command line and does nothing until you approve it.
Separately, any finished note, grab or translation can be posted to a URL you own, signed and timestamped. Secrets are encrypted on your machine, delivery retries in the background, and every attempt is shown in the Activity feed. Your endpoints are yours — they do not pass through us.
Before you ask
Questions people search for
Does Slack have dictation of its own?
In one place. Slack's Slackbot page describes a microphone icon for dictating a prompt to the assistant, says it works with the languages Slack supports and many others while being optimised for English, and says dictated messages are transcribed but audio data is not stored; where that audio is transcribed is not documented. It is the assistant's prompt box, not the message box you answer your team in. Deixis types into every box, on your own processor.
What about audio clips — is that not the same thing?
Different job. A clip is up to five minutes of you talking, posted for people to play; it is on all plans, it is not transcribed automatically, and someone has to press Generate transcript afterwards. Deixis gives you text in the message box, which is searchable, quotable and readable at a glance. Clips are lovely for a demo and slow for an answer.
Can Deixis take notes in a huddle?
It can record and transcribe one on your own machine, the way it records any call — Slack's own AI huddle notes are, by its help pages, a paid-plan feature that embeds the transcript in the notes canvas. Deixis's recording, transcription and transcript file are free and local, with no account. The difference is that yours stays on your machine and nobody else in the huddle gets it unless you send it.
Do I need an account or a subscription?
No. Dictation, both speech models, the dictionary, screen grabs, screen recording and meeting recording and transcription are free with no sign-in at all. The optional cloud cleanup is free too if you bring your own provider key, or point it at a model on your own machine. Paying only buys not having to set that up. Sharing a recording as a link is the one thing that needs a Deixis account.
Will it work in the desktop app and in the browser?
Both, and everywhere else. Slack ships desktop apps for Mac, Windows and Linux and mobile apps for iOS and Android; Deixis is a plugin for none of them, because it types at the caret in whatever has focus. The same key works in the Slack app, in a browser tab, in your editor, in your email and in your ticket tracker. What changes between them is the punctuation.
Will it work offline?
Completely. With no cleanup endpoint set, dictation, free mode, screen grabs and meeting recording and transcription all run with no network at all. Two calls a fresh install makes on its own: the update check, and — when you open the Settings window — a check for an announcement from us. Both send nothing but the request.
Keep reading
Related pages
Voice input for your product, without the audio leaving your user
Put voice input in your web or desktop app in about ten lines. Recognition runs on your user's machine, so no audio reaches you, us, or any cloud.
Offline dictation on Windows
Dictation that works with no internet: speech recognition on your own device, free, no account. Which Windows speech-to-text tools genuinely run offline.
Dictation for Microsoft Teams
Deixis dictates into Microsoft Teams on one key, and records and transcribes your calls on your own machine. No bot in the call, no account, free.
Dictation for Outlook
Deixis dictates into Outlook on one key — subject lines punctuated as subject lines, bodies as prose. Speech recognition on your device, no account.
Dictation for Claude Code
Deixis dictates into Claude Code on one key — terminal, editor, desktop app or browser. Speech recognition on your own device, no account, free.
Dictation for Cursor
Deixis dictates into Cursor on one key — the agent prompt, the editor, the terminal, anywhere else. Recognition on your own device, no account, free.
Hold a key. Speak. Keep working.
Deixis is free, needs no account, and runs on your own device. Everything on this page is in the current release.
v1.2.2 · 64‑bit Windows 10/11 · macOS 13 on Apple silicon · 51 MB · nothing else to install · also for Windows