1. Decide what the speech should become
A chat message is editable text that you review before sending. Live captions are a temporary reading surface for current speech. A Meet transcript is a retained record of spoken words. Take notes with Gemini creates a separate Google Docs summary with decisions and follow-ups.
Choose the artifact before enabling a microphone or meeting feature. Dictation is appropriate when your own words should become one ordinary message. It is not a substitute for capturing every participant.
- Write one message: dictate into the focused Meet chat composer.
- Follow the call: turn on live captions for your own view.
- Retain spoken words: start a Meet transcript when eligible.
- Create a summary: use Take notes with Gemini when the plan allows it.
2. Dictate an editable Meet chat message on a Mac
Open Chat in the intended meeting, click the message composer until the insertion point is visible, and trigger the Dictation shortcut configured on your Mac. Speak, stop Dictation, review the visible draft and audience, then click Send yourself.
Meet chat visibility depends on the meeting and account. Google documents continuous meeting chat for eligible Business and Enterprise users; when continuous chat is unavailable, messages may disappear when the call ends. External guests can have different access from internal invitees.
- Confirm the correct meeting and audience before speaking.
- Keep the draft unsent while checking names, dates, links, and tone.
- Treat Dictation, Meet microphone access, and the final Send action as separate controls.
3. Use live captions when the job is to read current speech
Google Meet live captions display subtitles as people speak. Google says captions are enabled for the individual viewer, and the viewer can scroll through caption text from the portion of the meeting they attended with captions turned on.
Google’s Meet privacy guidance says live-caption audio is temporarily sent to a Google transcription server, is not linked to user identifiers, and is not permanently stored. That boundary applies to live captions; it does not describe a saved transcript, recording, or notes document.
4. Use Transcripts only when spoken words should be retained
Meet Transcripts are available only on listed Workspace editions and supported devices. Host Management determines who can start them, and administrators may require participants to provide explicit consent.
Google says the transcript contains spoken words—not meeting chat—and is saved in the organizer’s Google Drive under the Google Meet folder. It is also attached to the Calendar event, with access depending on invitees, organization membership, and the meeting’s sharing rules.
5. Keep Gemini notes separate from a verbatim transcript
Take notes with Gemini creates organized meeting notes in Google Docs, can show a Summary so far, and can add suggested next steps. The feature requires an eligible Workspace subscription and currently supports one meeting language at a time.
The host chooses whether notes are shared with all invited guests, internal invitees, or hosts and co-hosts. If both note-taking and transcription are active, Google can add transcript citations to the notes. Review the generated document because summaries and action items can be incomplete or inaccurate.
6. Troubleshoot the exact path that failed
If no dictated text appears in Meet chat, test the same shortcut in TextEdit. If TextEdit also fails, inspect Keyboard → Dictation, the selected language, microphone input, and the current processing disclosure. If TextEdit works, refocus a plain Meet chat composer and try a short unsent draft.
If captions, Transcripts, or Gemini notes are missing, check the account edition, administrator policy, host role, meeting type, language, consent setting, Drive capacity, and whether the feature is enabled for the meeting. A working Mac microphone does not grant a Workspace feature or host permission.
7. Add IraVoice only for cross-app editable text
IraVoice 0.7.1 is a separate option for macOS 26+ Apple-silicon Macs when the same held-key workflow should produce editable text across supported Meet chat, email, documents, and developer fields. Speech recognition and supported formatting run on the Mac after model setup; paid continuation needs one Gumroad key check.
IraVoice does not join or record meetings, capture other participants, generate live captions, create Meet transcripts, summarize a call, alter sharing permissions, or press Send. It targets a supported focused editable field and falls back to the clipboard when insertion is uncertain.