How to Dictate in Notion and Gmail

Search "dictate in Notion" and you land on browser extensions, not a Notion feature — because Notion still doesn't let you talk into a page the way you talk into a phone's text field. Same story in Gmail: there's no microphone icon in the compose window. Two things actually work today. The first is a browser extension — most commonly one called Voice In — that adds a mic button to Notion, Gmail and thousands of other sites, running on your browser's own built-in speech engine. The second is a system-level dictation tool that runs underneath every app on your machine, so one hotkey works in Notion's desktop app, Gmail, Slack, or anywhere else your cursor lands. They solve the same problem at two different layers, and where your voice actually goes depends entirely on which one you pick.

Does Notion have a native dictation feature?

Not for writing. Notion shipped its first real voice input on April 6, 2026, but it's scoped to one specific thing: talking to Notion AI. The release note says it plainly: "Voice input for Notion AI is now on desktop. Whether you are dictating a longer prompt, or just prefer talking over typing, voice input makes it easier to get things done with your Agent." That's a microphone for prompting the assistant — not a way to speak a paragraph into a normal text block while you write.

Notion's most recent release as of this writing, 3.6 (July 1, 2026), adds the ability to upload a recorded meeting so AI Meeting Notes can transcribe it — useful, but that's file transcription after the fact, not live dictation into a page as you type. If what you actually want is to talk to Notion AI itself, that's the same idea we cover in our guide to voice prompting AI tools — Notion's new mic is an AI-prompt button, not a writing tool, and treating it like one is where most of the confusion comes from.

Does Gmail have built-in voice typing?

No. Gmail's compose window has never had a microphone of its own. Voice typing is a Google Docs feature (Tools → Voice typing), not a Gmail one, and the two get confused constantly because they're both Google products living in the same browser tab. If you want to speak an email in Gmail today, you have to reach for something outside Gmail itself — your OS's dictation, a browser extension, or a system-level dictation app.

What actually shows up when you search this: browser extensions

The most-recommended fix for Notion specifically is a Chrome/Edge extension called Voice In, and it's popular enough that people write full setup guides around it and swap tips for it on r/Notion. Dictanote, the company behind it, says the extension is "loved by 600,000+ people globally" — right-click a text field, hit start, and it transcribes in real time across Gmail, Notion, Slack, WordPress and reportedly thousands of other sites, in 100+ languages.

It has two real limits worth knowing before you install it. First, per Dictanote's own FAQ, it "only works within the browser" — it can't reach Notion's native desktop app, a native Outlook window, or Slack's desktop client, because those aren't web pages. Second, it's Chrome and Edge only, on desktop. Firefox ships the same underlying API disabled by default behind a flag, and support elsewhere is inconsistent — so this entire category of extension mostly doesn't exist for you outside a Chromium browser.

What does an extension like Voice In actually do with your voice?

This is the part that's easy to miss. Voice In doesn't run its own speech-recognition model — it's a thin layer over whatever speech engine your browser already ships. Dictanote says so directly in its own support docs: "Voice In uses the browser's speech-to-text API to transcribe text. On Chrome it uses Chrome's speech-to-text API, on Edge it uses Edge's API... No audio data or the resulting transcripts are sent to Dictanote's server." That's accurate, and it's a fair privacy claim — for Dictanote's own servers specifically.

But "transcribed by your browser" isn't the same claim as "transcribed on your device." The API in question is the Web Speech API, and its own documentation is explicit about how Chrome implements it: on-device speech recognition is opt-in, and by default, "using Speech Recognition on a web page involves a server-based recognition engine. Your audio is sent to a web service for recognition processing, so it won't work offline." A side-by-side comparison from AssemblyAI names the destination: Chrome "sends audio to Google's servers for processing." So every sentence you dictate into Notion or Gmail through this kind of extension leaves your machine, headed for Google's (or Microsoft's, on Edge) cloud speech service, before the transcript comes back to your page. Dictanote never touches it. Google or Microsoft does.

None of that makes Voice In a bad tool — it's upfront about the mechanism, free to use, and does exactly what it advertises. It just means "my dictation company doesn't see my audio" and "my audio never left my machine" are two different claims, and only the first one is actually true here.

The structural fix: one dictation tool, every app

The real limitation isn't Notion or Gmail specifically — it's that a browser extension is bound to the tab it's installed in. It can't reach Notion's desktop app, Slack's desktop app, or a native Outlook window, because none of those are web pages a content script can inject into. The fix that works everywhere at once is dictation that runs underneath all of it, at the operating-system level: one hotkey, and whatever app currently has focus receives the text — browser or not.

Inkvox works this way. Tap a key, talk, tap again, and the words land wherever your cursor is — Notion desktop, Gmail in any browser, Slack, Word, your terminal. The part that matters most for this comparison: Inkvox runs Whisper large-v3-turbo, quantized, directly on your own GPU through Vulkan (NVIDIA, AMD or Intel, with a CPU fallback if you don't have a compatible GPU) — not in a browser API, not on someone else's cloud. On a machine with an RTX 3070, a sentence transcribes in roughly 0.3–0.4 seconds. Audio uploaded: 0 bytes. No account required, and it's fully offline after the one-time ~800 MB model download — which also means it keeps working with zero signal, the same scenario we walk through in our guide to dictating fully offline. Inkvox supports 100+ languages and, like the browser API it's replacing, works in every text field you can click into — see how it works for the full mechanics.

How to dictate a Notion page, step by step

1. Open the Notion page — desktop app or browser, it doesn't matter — and click into the block where you want text.
2. Tap Right Ctrl (or whichever key you've assigned as your dictation hotkey).
3. Talk normally, the way you'd explain the paragraph to a coworker.
4. Tap the hotkey again — dictation stops and the transcript types straight into the block, at your cursor.
5. Add Notion formatting after the fact — toggles, callouts, headings — the same way you would with typed text. Dictation gives you the words; structure still comes from you.

How to dictate a Gmail email, step by step

1. Open Gmail and click Compose, then click into the message body.
2. Tap the hotkey and speak the email — say "period," "comma," or "new paragraph" out loud if you want that punctuation dictated explicitly, or just let it land as one block and edit after.
3. Tap the hotkey again to stop; the transcript appears exactly where your cursor was.
4. Read it back once before sending — same habit as any dictated text, a ten-second pass catches the rare misheard word.

If you'd rather not install anything system-level and only need this inside Chrome or Edge, Voice In covers the same steps in the browser — just remember it's routing through Google's or Microsoft's cloud speech service to get there, not Dictanote's, and not your machine.

FAQ

Does Notion have a native dictation feature?

Not for writing. Notion added voice input on April 6, 2026, but it's scoped to prompting Notion AI/Agent, not to speaking text into a normal page block. As of Notion 3.6 (July 1, 2026), the only other voice-related addition is uploading recorded audio to AI Meeting Notes for after-the-fact transcription — still not live dictation while you write.

Is the Voice In Chrome extension private?

Partly. Voice In itself doesn't send your audio to its maker's (Dictanote's) servers — it hands the microphone to your browser's own built-in Web Speech API. On Chrome and Edge, that API is server-based: it sends the audio to Google's (or Microsoft's) cloud for transcription. So no third party you didn't expect sees it, but it isn't processed on your device either.

Can I dictate directly in the Notion desktop app?

Not with a browser extension — Voice In and similar tools only work inside the browser tab they're installed in, so they can't reach Notion's native desktop app. A system-level dictation tool like Inkvox works underneath every app, including the desktop client, because it types wherever your cursor is rather than injecting into a web page.

Does Gmail have built-in voice typing?

No. Gmail's compose window has never had a microphone button — that feature exists in Google Docs (Tools → Voice typing), not in Gmail. To dictate an email you need something external: your OS's own dictation, a browser extension, or a system-level dictation app.

The gap in Notion and Gmail was never really about those two apps — it's that dictation built into a browser tab can only ever cover browser tabs. Join the Inkvox waitlist to dictate into Notion, Gmail, and everything else, without your voice making a stop at someone else's server first.