How to Transcribe a Zoom Recording
Get a transcript of a Zoom meeting: what Zoom's own transcript gives you, where the recording file actually is, and when to transcribe it yourself.
The Scriptorium team ·
Zoom can already give you a transcript, and if the meeting was recorded to the cloud you may not need anything else. It is worth knowing what that transcript is before you decide — because for a lot of meetings it is not quite the thing people want.
First: check whether Zoom already made one
Cloud recordings can be transcribed automatically by Zoom itself. If your account has it switched on, the transcript is waiting for you:
- Sign in to Zoom on the web and open Recordings.
- Open the meeting you want.
- If a transcript exists, it appears beside the recording as Audio Transcript.
What you get is a VTT file — a caption file, with a timestamp on each short chunk of speech. That is genuinely useful for captions. It is less pleasant to read as a document, it has no speaker labels beyond what Zoom guessed, and you cannot correct it and have the correction follow you anywhere.
This only exists for cloud recordings, on paid plans, when transcription was enabled before the meeting. A local recording has no transcript, and enabling the setting afterwards does not go back and make one.
Where the recording file actually is
If you recorded locally — the default on a free plan — the files are on the computer of whoever pressed record, not in the cloud. Zoom writes them when the meeting ends, and until that conversion finishes there is nothing to find.
- Windows:
Documents\Zoom, in a folder named for the date and meeting. - macOS:
~/Documents/Zoom, same arrangement. - In the app: Settings → Recording shows the exact location, and Meetings → Recorded lists them with an Open button.
Inside that folder you will usually find zoom_0.mp4 — the video — and audio_only.m4a. Either will transcribe. The audio file is much smaller and produces exactly the same text, so prefer it if you have it.
Transcribing it yourself
If you want a readable transcript with speaker turns, one you can correct and export in a format that is not a caption file, upload the recording:
- Open video to text — or audio to text if you are using the
audio_only.m4afile. - Drop in the Zoom file. There is no conversion step first; MP4 and M4A both go straight in.
- Wait for the transcript. Transcription runs faster than real time, so a one-hour call is usually done in a fraction of an hour.
- Read it against the audio and fix the names — meeting recordings are full of proper nouns, and proper nouns are what automatic transcription gets wrong most.
- Export as plain text or Markdown for notes, or SRT/VTT if you did want captions after all.
Scriptorium cannot reach into your Zoom account, and there is no Zoom integration. It transcribes files you upload. That is a real limitation and it is better said here than discovered after you have gone looking for a connect button.
Which route to pick
| What you need | Use |
|---|---|
| Captions to upload back to a video | Zoom's own VTT, if it exists |
| A readable record of who said what | Upload the recording and transcribe it |
| To correct mistakes and keep the corrected version | Upload it — the Zoom transcript is not editable in place |
| A transcript of a local recording | Upload it — Zoom never made one |
| To search across many past meetings at once | Upload them to one account |
Getting a better recording next time
Almost everything that decides transcript quality is decided while recording, not afterwards. On a call that means: ask people to use a headset rather than a laptop microphone, ask them not to talk over each other, and record the meeting rather than recording the room the meeting is playing in. A call recorded through a laptop speaker and re-captured by a phone is the worst case there is.