- AI summary
- Built for Google Meet
Get the short version of a Google Meet call without keeping the audio
One click after the call turns the transcript into topics, decisions and next steps, readable in either of the two languages that were spoken — and the audio it came from was never stored.
The AI meeting summary is included in every paid plan. It is not part of the 30-minute trial.
32 languages, any pair, chosen before you join.
The call this page was written for
Six calls in a day, three of them in a language you use only for work, and by the evening they have merged into one long blur. Rewatching is not an option, because there is nothing to rewatch: you never recorded anything, deliberately. The summary gives you the recap you would have written at the time, without a recording file existing anywhere to be leaked, requested or quietly forgotten about.
What this looks like during the call
Live transcript
ESENA one-click recap of topics, decisions and next steps, readable in either meeting language.
El proyecto piloto duró dos semanas más, así que las cifras finales solo llegan el jueves.
The pilot project ran two weeks longer, so the final numbers only land on Thursday.
Entonces movamos la revisión al viernes por la mañana. Hoy aviso al cliente.
Then let's move the review to Friday morning. I'll let the client know today.
Me sirve. Mándame el borrador la noche anterior, con el desglose de costes.
Works for me. Send me the draft the evening before, with the cost breakdown.
Action items
Built for Google Meet- Move the review to Friday morning
- Tell the client today that the numbers land Thursday
- Send the draft with the cost breakdown on Thursday evening
What usually goes wrong
- Captions alone do not give you a reusable bilingual document; check the recording and export options before relying on them.
- A recording is the easiest thing to make and the hardest thing to still be responsible for six months later.
- Notes written after three back-to-back calls describe the last one and guess at the other two.
Why this holds up in a real call
- The recap is generated from the call's own transcript; no recording file is kept to produce it.
- The recap is produced from the call's own transcript, and the audio behind it is processed in transit rather than written to disk on our side.
- Each side reads the recap in its own language, so nothing has to be re-translated before it is forwarded.
How the three options compare
What Google Meet does on its own
Caption display alone does not generate a summary. Where meeting recording is available, Record captions can include captions in the recording. Turning on captions alone does not automatically create a bilingual PDF, SRT or TXT export; availability and permissions depend on the plan and settings.
Standard captions transcribe speech; translated captions display translated text on eligible plans. Both are text features. Speech Translation is a separate feature for translated audio. Google announced general availability of Speech Translation on February 4, 2026. Current Help lists English paired with French, German, Hindi, Italian, Portuguese or Spanish: one pair per meeting, with a 90-minute limit. Access depends on an eligible plan, administrator settings, supported devices and participants allowing their voices to be translated. Speech Translation is not available in recordings or live streams. Where meeting recording is available, Record captions can include captions in the recording. Turning on captions alone does not automatically create a bilingual PDF, SRT or TXT export; availability and permissions depend on the plan and settings.
| What you are comparing | MeetVoice | Built-in Google Meet | Notetaker bot that joins the call |
|---|---|---|---|
| How fast do you get a readable recap? | One click after the call, built from the transcript that was already being written. | Caption display alone does not generate a summary. | After the vendor has finished processing the meeting on its own servers. |
| Does a recording have to exist first? | No. Audio is processed in transit and no recording file is produced. | No, but there is also no recap. | Usually yes — most keep the recording as the artifact the summary is built from. |
| Who else holds a copy? | Nobody. The transcript and the summary stay in your own history. | Nobody, unless your workspace has recording or transcripts switched on. | The vendor, for as long as its retention policy says. |
| What does it take to get started? | Any paid plan; the 30-minute trial covers translation but not the summary. | Nothing beyond the workspace plan your organization already has. | An account, plus permission for the bot to be admitted to your calls. |
How to set it up
MeetVoice is a Chrome extension plus a companion desktop app. The extension reads the Google Meet tab you already have open and the desktop app coordinates the session from your machine, so no extra participant is added to the call.
Step 1 Install MeetVoice and activate a paid plan
Add the extension from the Chrome Web Store, install the desktop app, and activate any paid plan. The 30-minute trial covers live translation and transcripts; the summary is not part of the trial.
Step 2 Run the call as you normally would
Start the session at the beginning of the call. Subtitles and voiceover do their job while the transcript builds quietly in the panel, and no recording file is created at any point.
Step 3 Ask for the summary afterwards
When the call ends, request the summary from the panel. Key topics, decisions and action items appear in both meeting languages, and the full transcript is still there to export separately.
Questions people ask before installing
Is the summary included in the 30-minute trial?
No. The trial covers live subtitles, voiceover and the transcript. The summary comes with paid plans, so a trial account will not show it.
If nothing is recorded, what is the summary made from?
From the text transcript MeetVoice writes while the call runs. The audio itself is processed in transit and is not stored.
Can each side read it in their own language?
Yes. The summary is produced in both meeting languages from the same transcript, which is why the two versions describe the same decisions instead of drifting apart.
Do I need my own speech-recognition key for this?
No. Speech recognition is included in the trial and in every plan, so there is no key to obtain or connect. The summary is built from the same live transcript either way.
Get the short version of a call you have already had, without rewatching anything.
The AI meeting summary is included in every paid plan. It is not part of the 30-minute trial.
Page last reviewed