How to record and transcribe a Google Meet call
2026-07-19
Google Meet only records calls on paid Workspace plans, and even then you get a video file rather than anything you can read. If what you actually want is the text — who agreed to what, what the client asked for, what you promised to send — you need two things: a recording, and a transcript with speaker labels.
Both are free. Here is the whole workflow.
1. Record the call
Install Meet Voice Recorder. It captures the tab audio — that is, everyone else in the call — and mixes in your microphone, so you end up with one file containing the whole conversation.
Open your Meet call and press Ctrl+Shift+Y to start. Press it again to
stop. The recording lands in your Downloads folder as
meet-YYYY-MM-DD-HHMM.webm.
The extension does everything locally. Nothing is uploaded, there is no account, and it keeps no copy — the file it hands you is the only one that exists.
2. Turn it into text
Drop that .webm onto the whosayah home page. It is accepted
as-is; there is no need to convert it to MP3 or extract the audio first.
Once it has uploaded, a window shows what we found — the file’s name, size and length — and asks how to handle it.
Leave Label speakers switched on there. Without it you get one undifferentiated wall of text, which for a multi-person call is nearly useless. With it, each line is attributed, so you can actually follow who said what. Enter how many people were on the call, then press Transcribe. If you genuinely do not know, I don’t know — detect it starts it without a count.
3. Read, or download
Lines appear as they are transcribed, so you can start reading before the job finishes. When it is done you can copy the whole thing, or download it two ways:
- .txt — a plain transcript, one line per turn, speaker prefixed. This is the one you paste into notes.
- .vtt — WebVTT subtitles with timestamps, if you also recorded video and want captions.
Voices arrive as Speaker 1, Speaker 2 and so on, because the software can tell them apart but cannot know their names. Press Name speakers to put real names to them — each voice is shown with how much it talks and its first line, which on a call you have just sat through is usually enough. The names go into both downloads, not just the page.
How good are the speaker labels?
Good on a normal call where people take turns and the audio is clear. Two things make it harder: heavy crosstalk, and very short interjections. A one-word "yeah" carries little voice information, so those are the lines most likely to be attributed to the wrong person. Longer turns are reliable.
If you know how many people were on the call, tell us: the How many people are speaking? box in that window is the single field that most improves the result. Left to guess, we tend to split one person across several speakers.
A note on privacy
Recording a call usually means recording other people. Tell them. In many places it is a legal requirement, and everywhere else it is simply the decent thing to do.
On our side: the uploaded file is deleted as soon as the transcript is ready, and the transcript itself is deleted after 24 hours. There is no account, so there is nothing tying either to you.