Can you turn on transcription in Google Meet?

You went looking for one toggle and found a settings maze; that is a fair description of how Google Meet splits this feature. This section covers where transcription sits, who is allowed to start it, and what happens the second it turns on. In short: it is in the Activities panel, it is gated by the host account's Google Workspace edition, and it announces itself to the room.

• Where the button is: the Activities icon in the meeting bar, then Transcripts, then Start transcription.
• Who can start it: the meeting host, plus participants inside the host's organization when the host has not restricted it.
• What you end up with: a Google Doc written into the organizer's Drive after the call ends.

Personal Gmail accounts do not have Transcripts at all. They have live captions, which render text on your own screen during the call and disappear when it closes. The wider set of Google Meet interview questions is collected under Interview Types.

Live captions are visual and local: Meet draws text under the video tiles for the person who switched them on, and keeps no copy. Transcripts are a saved document with speaker labels, and they are the version that recruiters, legal teams, and hiring notes actually reference.

The difference matters in an interview. Captions help you follow an interviewer with an unfamiliar accent or a rough connection, and nobody else learns you turned them on. A transcript is a shared artifact: the room is told it started, and the file outlives the call by however long the organizer keeps it.

If your reason for wanting a transcript is comprehension rather than record-keeping, captions are the better answer and cost nothing.

A transcript is a record of what was said. It arrives after the call, it preserves your pauses and your weakest answer word for word, and it does nothing for you while the interviewer is waiting.

Consider a backend engineer taking a senior round with a payments company over Google Meet. She starts transcription hoping it will support her, and the only thing that changes is that the hiring manager now has a searchable copy of the moment she blanked on idempotency keys. What she needed was a private prompt during the pause, not a document afterwards.

That gap is why SubcueAI transcribes locally instead of asking Meet to do it. The desktop app on macOS and Windows captures system audio and your microphone together, so both sides of the conversation become text on your own machine; the Chrome and Edge Side Panel captures the meeting tab's audio, which is the interviewer's side, for browser-tab calls. Nothing is switched on inside Google Meet, no bot joins the participant list, and nothing is injected into the meeting page. The pipeline is broken down under How It Works.

Neither a Meet transcript nor a local assistant changes what other people can observe. The boundaries worth holding on to:

• Screen sharing: anything on the shared display is visible, including an assistant overlay if you share a whole screen instead of a single window.
• Recording: a recorded call keeps whatever the camera and the shared screen showed, for as long as the recording exists.
• Proctored assessments and company-managed devices: these run their own monitoring, and no third-party tool should be assumed safe inside them.
• Consent rules: recording and transcription law varies by country and by state, and a transcript is a durable record of the call.

Those limits are covered across Detectability and Privacy, and how audio and transcripts are handled is written up on the /security page.