How to transcribe a Google Meet call for free (no bot)

October 4, 2026·Updated November 18, 2026·guide·works on any Meet call, any plan

Meet's own recording is a paid Workspace feature that announces itself to everyone and lands the file in Google Drive. Third-party bots like Otter join as a visible participant. There's a third path: capture the meeting tab's audio with a browser extension and transcribe it on your own machine. This guide uses QuickSpeak (which we make), and the whole flow takes under a minute to set up.

Before you start

Install QuickSpeak from the Chrome Web Store. There's no account and nothing to configure. On first use, the app downloads its speech-recognition model once and caches it — after that, recording and transcription work without any network at all.

The steps

  1. Join your Meet call in a Chrome tab as usual. Nothing about the meeting changes — no bot appears in the participant list, because nothing joins.
  2. Open QuickSpeak from the toolbar and set Audio source to Browser tab (Meet, Zoom web). That's the default for tab-based calls.
  3. Optional: flip on Record video if you also want the visual record.
  4. Click Start recording and close the popup. QuickSpeak sits quietly in your toolbar while you pay attention to the meeting.
  5. Click Stop recording when the call ends. You'll see three stages — Transcribing, Identifying speakers, Saving — with the standing hint: "Working locally — nothing leaves your device."
  6. Name your speakers. The transcript labels voices "Speaker 1, Speaker 2…". Click any label and type the real name; it updates throughout that transcript. Recurring Meet vocabulary — client names, project terms — belongs in custom vocabulary, so future calls transcribe it correctly from the first pass.
  7. Export. Copy transcript pastes it anywhere; download it as .txt. Pro adds Markdown, JSON, and WAV — and during your first 14 days, everything Pro is unlocked, no card required.

Meet's native transcription vs a local transcript (November 2026)

Verified against Google's Workspace feature comparison this month: saved transcription ("Transcribe meetings and save them to Drive") is available on Business Standard, Business Plus, and Enterprise editions — not Business Starter, and not free personal Google accounts. Personal users get live captions during the call, and AI note-taking only with a Google AI Pro or Ultra subscription. The distinction matters: captions (on-screen text) and transcription (the saved record) are different features, and the caption button being visible doesn't mean a transcript is being kept.

Even where the edition qualifies, the transcript belongs to the organizer's organization — saved to their Drive, governed by their admin's retention policy. Guests, external clients, and personal-account participants can't generate their own. The full edition-by-edition comparison is in QuickSpeak vs Meet's built-in transcription; the short version is that a local transcript is the only kind that belongs to the attendee. And if you do have Workspace transcription, the two aren't rivals: Meet's saved transcript serves the team, while a local copy serves your own archive.

Why this works on any Meet call

Google gates its native recording behind Workspace editions and announces it to the whole call because Google's recording happens in Google's cloud — it's Google's file, stored in Drive, governed by Google's retention. A recording bot raises the same flag from another direction: it's a participant, so participants see it, and some organizations block it.

Capturing the tab's audio sidesteps both. The audio your browser already plays is recorded on your machine, and the recognition model runs in WebAssembly in that same browser. There's no upload step to announce, and no host permission to request. Your transcript is a file on your disk.

The consent caveat: a tool that doesn't announce itself doesn't change the law. Depending on where your participants are, recording may require one party's consent or everyone's. Disclose it — "I'm recording for notes" costs nothing.

What the free tier includes

Unlimited recording and transcription — there is no minutes meter, because transcription runs on your hardware. Speaker identification per meeting, full meeting history, copy and .txt export, and audio export are all free. Pro ($12/month or $99/year) adds live captions during the call, search across all your past meetings, speaker profiles that persist across meetings, and more export formats. Every install starts with 14 days of full Pro, no card required — long enough to transcribe real meetings before deciding.

FAQ

Will Meet participants know I'm transcribing?

No notification appears and nothing joins the call — QuickSpeak records the tab's audio on your own machine. Disclosure to participants is still your responsibility under your local consent laws.

Is it really free?

Recording and transcription are unlimited on the free tier, with per-meeting speaker ID and .txt/audio export. Pro adds live captions, cross-meeting search, and Markdown/JSON/WAV export, with a 14-day free trial on every install.

How is this different from Meet's built-in recording?

Meet's recording needs a paid Workspace edition, host permission, announces itself, and stores the file in Google Drive. This works on any call, needs nobody's permission, and the transcript stays on your disk.

Where is the transcript saved?

In your browser profile, on your disk. Transcription is WebAssembly running locally — there's no server involved, and you can export or delete the file whenever you like.

Does it work with Zoom and Teams too?

Yes — same flow when the call runs in a browser tab. For the Zoom or Teams desktop apps, switch the audio source to Desktop (Zoom app, Teams). See our Zoom guide and Teams guide.

Sources & notes

  1. QuickSpeak features, pricing, and limits cited in this article: the shipping extension and usequickspeak.fyi, accessed November 17, 2026.