Transcribe a Google Meet Recording Straight from the Drive File

Where Google Meet keeps recordings

Recording in Google Meet is a Google Workspace feature. When the organiser or a participant from the same organisation presses Record, Meet processes the video after the call and saves it as an MP4 into the organiser's Google Drive, in a folder named Meet Recordings at the top level of My Drive. The organiser also receives an email with a link to the file. Nothing is saved on anyone's computer, and participants outside the organisation receive nothing unless the organiser shares the file.

Transcrify cannot read a Drive link, because the file sits behind the owner's permissions. The route is always the same: open the file in Drive, download it, upload it here. If you can play the recording in your browser but the Download option is greyed out, the owner has disabled downloading, and you will need to ask them for a copy.

Who ends up needing this

The most common case is the participant who was not the organiser. Meet's own transcript, where it exists, is delivered to the organiser alongside the recording, and it reaches the rest of the group only if that person forwards it. A customer who sat through a vendor demo, a candidate who wants to review an interview, or a student who joined a lecture from another school all have the same problem: the recording exists, but the text does not.

The second case is the free Gmail account. Meet does not offer recording there at all, so people capture the call with a screen recorder on their laptop or phone instead. Those files, usually MP4 or MOV, upload here just as well as a Drive recording. The third case is a recording where the organiser never switched on transcription, or where the transcript came out in the wrong language because the account default did not match the language spoken.

How the transcript handles a Meet call

Audio from Meet is clean by the standards of recorded speech. Each participant has a dedicated microphone, the platform filters keyboard noise and echo, and people generally wait their turn. The model assigns a label to every voice it can tell apart and keeps that label consistent through the recording, so the output reads as a dialogue with timestamps rather than one long paragraph. The language is detected from the opening seconds, and more than fifty languages are supported, which covers a call that starts in English and switches to Spanish halfway through.

An hour of audio takes about three to five minutes, because the recording is processed in parallel segments rather than from start to finish. The free limit is 3 hours per file, which is enough for a half-day training session. The recording is processed on our own servers and is not forwarded to a third-party speech API, so a sales call or a performance review stays between you and the people who were on it.

What a free account adds

The guest transcript can be read and copied. With a free account you can edit the text, which is useful for product names and the surnames of people on the call, keep a history of recordings, and download the result as TXT, SRT, VTT or ASS. The account includes 5 hours across 10 files, and the transcripts you created as a guest are moved into it on sign-up. AI chat with the recording, automatic summaries and export to PDF or Word are part of the paid plans.