Voice dictation for Obsidian
A vault fills up with words, and most of them are yours. A meeting you want to remember, a decision you want the reasoning for, the paragraph explaining why the schema looks like that. Typing is a poor way to catch a thought you have already had — it is slower than thinking, and by the time the sentence is finished the second half of the idea has gone. Talking is the fast way to get a thought into a note, and a note is prose, which is the one thing dictation is genuinely good at.
Deixis is voice input for exactly that. Hold Ctrl+Win, say the paragraph, let go, and clean
punctuated text is typed straight into the note you were looking at, with your clipboard handed back
the way you left it.
And this is the part that matters: Deixis is not a plugin, and not an extension for anything. There is nothing to install into Obsidian, nothing in your community-plugin list, and nothing to keep up to date when either app changes. It types wherever your cursor already is, so the one key dictates your note, your email, your ticket, your chat and your browser's address bar. One voice input for the whole machine, not one per app — on Windows, Mac, iPhone and Android, with speech recognition running on the device every time and audio that never leaves it.
Why Deixis
What it gives you in Obsidian, and what it will not do
Why Deixis in Obsidian
- Everything it makes is already a vault note. Obsidian's own help calls a vault "a folder on your file system which contains your notes, attachments, and the configuration folder". Deixis writes to a folder too: a free-mode session comes out as one Markdown file with its images in a folder beside it. There is no export step, because there is no proprietary format to export from.
- Talk and point at the same time. Double-tap the key and Deixis keeps listening while you drag a region of the screen, pick a window element, or grab a text selection. What you said and what you pointed at come out as one note that interleaves the two — narration and evidence, in order. That is a research note, a bug report or a design review dictated in the time it takes to describe it, and it lands as Markdown with the images sitting next to it.
- Screen grabs, annotated, saved as ordinary PNGs.
Ctrl+Win+Sfreezes the screen, you drag a region, and you either quick-save it or open the editor — arrow, box, pen, highlighter, text, numbered step badges and a pixelate tool for redacting. Every save keeps a sidecar so the grab stays re-editable rather than flattened forever, and Obsidian's file-formats page is happy to embed the result. - A screen recording, with a link, for what a note cannot show.
Ctrl+Win+Vrecords the screen you are on to an MP4 with a small bar showing the elapsed time. If a microphone was recorded, Deixis transcribes it on your own machine and writes the recording up with a title and a summary, with the transcript beside the video as Markdown. Sharing it as a link is the one part that needs a Deixis account: whoever opens it needs none, you set how long it lives, and you can revoke it. - Meetings, transcribed on this machine.
Ctrl+Win+Mrecords your microphone and the system output and transcribes both as they run, with no key, no account and no network. 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. The audio and the transcript sit in a folder you own, and the transcript is Markdown — which is to say, already a note. - Your own buttons, which is how a note reaches the vault by itself. A button on the palette can run a program with the finished note as JSON and type back whatever it prints, straight at the caret — so a short script of yours can file the note into the right folder of your vault and hand you the link. Steps can also be chained as plain settings — translate, summarise, apply an instruction of your own, post to a webhook — with no scripting language at all, and the panel has ready-made ones a click adds.
- Signed webhooks 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. Images travel as paths, which is exactly right for a receiver on the same machine: it can move the files into the vault itself. Delivery retries in the background and never makes you wait.
- It learns the words your notes are actually about. Library names, product names, the people on
your team. Select a word anywhere and press
Ctrl+Win+Dand it is in your vocabulary, fixed on this machine before anything is sent anywhere. Five packs ship in the box — software, accounting, business, medical, legal — and the software one is on by default. - Free, with no account at all. Dictation, both speech models, the dictionary, free mode, meetings, screen grabs and recordings, your own buttons and webhooks: all free, none of it behind a sign-in. Deixis never asks you to make an account to dictate.
What it will not do
- There is no Obsidian plugin, and it does not know your vault. Deixis types at the caret and
stops there. It does not read your folder structure, resolve a
[[wikilink]], know your tags or your properties, and it will not file a note for you unless you write the button that does it. If you want a vault-aware tool, this is not one. - It does not read your screen. Deixis is dictation, not an agent. It sends the name of the app and the name of the field along with your words so the punctuation fits, and never their contents. It cannot see the note you have open, so you still have to say what you mean.
- English only. The speech models are English models. Obsidian is happy in any language you can type; Deixis is not, and if you write your notes in another one this is not your tool yet.
- It will not transcribe a file you hand it. Deixis transcribes what it records — your dictation, your meetings, your screen recordings. If you have a folder of voice memos already, including ones Obsidian's own Audio recorder left in your vault, Deixis will not turn them into text.
- No Linux build, and no PC older than 2013. Obsidian ships for Linux and Deixis does not, and Linux is not planned. The speech engine is compiled against an instruction set that arrived in 2013, so an older processor stops with an error rather than running slowly.
- It is not a keyboard replacement. Brackets, links, tags and the
- [ ]in front of a task are still faster typed. Deixis is for the prose half of a note, which is 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 Obsidian
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 your vault, in the editor, kind by kind.
| Where you are | What you say | What Deixis types |
|---|---|---|
| A daily-note entrytoday's note, opened by the Daily notes core plugin | “um spent the morning on the migration script it choked on the null created at column so I backfilled it before rerunning” | Spent the morning on the migration script. It choked on the null created_at column, so I backfilled it before rerunning.The filler goes, the run-on becomes two sentences, and your dictionary puts the underscore back in a column name. |
| A long paragraph in a notethe body of a note in the editor | “the argument for local first is basically that a note you can read in a text editor in ten years is worth more than one that needs a running service uh and every sync feature has to be built on top of that not underneath it” | The argument for local-first is that a note you can read in a text editor in ten years is worth more than one that needs a running service. Every sync feature has to be built on top of that, not underneath it.A long field gets sentences and a hyphen where the sense needs one. This is the field where talking beats typing by the most. |
| A task linea checkbox line in a note | “ask the design team for the new empty state before friday” | Ask the design team for the new empty state before Friday.A short line is punctuated as a short line and nothing is expanded. The `- [ ]` in front of it is yours to type. |
| A note titlethe file name field when you make a new note | “notes on the tan stack upgrade” | Notes on the TanStack upgradeOne dictionary line fixes the spelling of a library your speech model has never met, on this machine, before anything is sent anywhere. |
| A card on a canvasa text card on a canvas | “this box is the bit that reads the feed twice on a cold start” | This box is the bit that reads the feed twice on a cold start.A card is a small field, so it gets a capital and a full stop and is left alone otherwise. |
Obsidian facts from home, download, its site, its privacy policy, vaults, formats, syntax, plugins, audio, properties and canvas, read on 7 September 2026. Everything about Deixis on this page is from its README for the current release.
Worth knowing about Obsidian
- Obsidian's own help page defines a vault as "a folder on your file system which contains your notes, attachments, and the configuration folder with Obsidian-specific settings" — which is the whole reason a Deixis note drops into one so easily: both sides are already just files in a folder.
- Obsidian's home page says it "stores your notes locally as plain text Markdown files", that you can reach them "even offline", and that "Obsidian uses open file formats, so you're never locked in." Free mode writes exactly that: one Markdown note with its images in a folder beside it.
- Obsidian does have a voice feature of its own: the Audio recorder core plugin, which its page says "lets you record and save audio in an Obsidian note", for the times "where you can't type fast enough". It "saves the recorded audio file to your vault and embeds it at the end of the active note", and the recording stays in the vault even after you remove it from the note.
- What that page does not describe is any transcription: it says nothing about speech-to-text, and no core plugin in Obsidian's own list is a dictation feature. So Obsidian records your voice as an attachment to listen back to; turning it into words in the note is not documented as something Obsidian itself does.
- The file formats page lists Markdown (`.md`), Bases (`.base`) and JSON Canvas (`.canvas`), plus images, audio, video and PDFs that "can be embedded directly into your notes" — so a dictated note and its screen grabs are both ordinary vault citizens.
- Obsidian's Markdown is the plain kind: headings from `#`, task lists from a hyphen and `[ ]`, and internal links as `[[Wikilinks]]` or ordinary Markdown links. Deixis types text at your caret, so which of those you are inside is your business, not its.
- Properties — the structured fields at the top of a note — are "stored in YAML format at the top of the file", holding text, lists, numbers, checkboxes, dates and tags. Dictation lands in one the same as in any other box.
- Canvas is described as a core plugin for visual note-taking with "infinite space to lay out notes and connect them to other notes, attachments, and web pages", saved as "`.canvas` files using the open JSON Canvas format".
- Obsidian's privacy policy is short and good. Of the desktop and mobile apps it says "All data is saved locally on your device and is never sent to our servers" and "We do not collect any telemetry data" — with Sync and Publish covered separately, as the paid services that do send data. Deixis makes the same promise about your audio, which never leaves the machine under any setting.
- The app is "Free without limits. No sign-up required." Sync is $4 per user per month billed annually, Publish $8 per site per month billed annually, and a commercial licence is $50 per user per year.
- Obsidian's download page offers desktop builds for Windows, Mac and Linux, and mobile apps on the App Store and Google Play. Deixis covers four of those five: Windows, Mac, iPhone and Android, with no Linux build.
Five minutes
Setting it up
About five minutes, and nothing goes into Obsidian at all — no plugin, no settings inside it, no restart.
Install, and one download
One installer, and nothing else to install with it. The first launch shows a single Download button for the speech model — about the size of a short video, fetched once, and checked against a published size and hash before it is kept. From then on you are offline unless you decide otherwise.
The gesture
Hold Ctrl+Win, speak, release. That is the whole product. Double-tap it instead and you are in
free mode: it keeps listening until you double-tap again, then pastes the finished note where you
point and saves it as Markdown. Free mode is the one to use for a thought you have not finished
having.
Hold the chord a moment longer and a small palette appears above the listening pill with a letter
per mode: S grabs and annotates a region, V records the screen, M records a meeting, T
translates what you have selected, D puts the word under your cursor into your dictionary. The
letters stay live for a few seconds and the palette disappears exactly when they stop working. If
Ctrl+Win fights something else in your setup, Settings captures a new chord as you press it.
Teach it your words
The biggest single improvement, and it costs nothing in speed. Select a word — a library in a
README, a name in a note — press Ctrl+Win+D, and it is yours. Names it keeps mishearing get a
correction that lands verbatim every time, and there is no limit on those. The Dictionary tab is
the same list if you would rather edit it in one place.
Decide what may leave, once
Settings → Cleanup is the only privacy decision here, and the 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 language model running on your own machine and even the polished prose stays local. Whatever you pick, the audio is not part of the choice — that never leaves, and there is no setting that makes it.
Getting notes into the vault
Two ways, and both are yours. A custom button runs a program you wrote with the finished note as
JSON: three lines that write the Markdown into a folder of your vault and copy the images beside it,
and stdout set to insert will type whatever your script prints back at your caret. The first run
shows you the exact command line and does nothing until you approve it, and editing the command asks
again. Or add a webhook and let a small receiver do the filing — it gets the note's text and
Markdown plus the paths of every image, signed, with retries in the background. Nothing in either
route passes through us.
Before you ask
Questions people search for
Does Obsidian have voice dictation of its own?
Not dictation, but it does have a voice feature: the Audio recorder core plugin. Its own page says it "lets you record and save audio in an Obsidian note" and that it "saves the recorded audio file to your vault and embeds it at the end of the active note" — so you get an attachment to listen back to, not words in the note. Nothing on that page describes transcription or speech-to-text, and no other core plugin in Obsidian's list is a dictation feature.
Do I have to install anything into Obsidian?
No, and that is deliberate. Deixis is a separate app that types at the caret of whatever has focus, so there is no plugin in your list, no version to keep in step, and nothing that breaks when either app updates. The same key works in your notes, your email, your terminal and your browser.
Where does the audio go?
Nowhere. Speech recognition runs inside Deixis on your own processor, and the recording, the model and the transcript stay on the machine — with no account and no telemetry. That matches what Obsidian promises about your notes: its privacy policy says that in the desktop and mobile apps all data is saved locally on your device and is never sent to its servers.
Can I dictate on my phone into my vault?
Deixis has an iPhone keyboard and an Android microphone bubble, and both dictate into any app's text field with the speech model running on the phone. Obsidian is free on iOS and Android, so the field you are typing into can be a note. What Deixis does not do is know it is a vault — it types the words and stops there.
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.
Do I need an account or a subscription?
No. Everything that runs on your machine is free with no sign-in: dictation, both speech models, the dictionary, meetings, grabs, recordings, buttons and webhooks. 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.
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 Jira
Deixis dictates Jira tickets on one key. Narrate a bug while you grab the screen, get one note with the picture in it, then file it from a button.
Dictation for Slack
Deixis dictates into Slack on one key — channel replies, threads, DMs and canvases. Speech recognition on your own device, no account, free.
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.
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
Facts about Obsidian checked on 2026-09-07 against obsidian.md. Deixis facts are from its README for the current release. Something out of date? Tell us.