Transcribe a Zoom Recording Without Uploading It

Transcribe a meeting recording you already have

Most meeting transcription products want to be in the meeting. They add a bot to the call, record from inside it, and store everything in their cloud, which means a subscription, an account for the workspace, and a visible extra participant that somebody usually has to explain.

If the meeting is already over and you have the recording on your disk, none of that is necessary. The tool above reads the MP4 or M4A that Zoom, Teams, or Meet saved and runs OpenAI's Whisper on your own machine to produce a timestamped transcript. No bot, no account, no plan tier, and the recording never leaves your computer.

Finding the file, then transcribing it

1. Zoom local recordings land in a Documents/Zoom folder, one subfolder per meeting. Teams and Meet recordings usually arrive via OneDrive or Drive and need downloading first.
2. Prefer the audio only file if there is one. It carries everything the transcript needs at a fraction of the size.
3. Drop it onto the panel above. The file is read locally, not uploaded.
4. Export plain text for notes and search, or SRT if you are going to cut the recording into clips later.

The consent problem an on-device transcript avoids

Recording a call is one decision. Sending that recording to a third party is a second, separate decision, and it is the one people tend not to make consciously. Everyone in the room agreed to being recorded. Whether they agreed to their voice being processed and retained by a transcription vendor is a different question, and in regulated work it is not a rhetorical one.

Running the model locally makes that second decision disappear rather than answering it. There is no vendor, no retention window, and no data processing agreement to go and read, because the audio only ever existed on the machine that recorded it and the machine you are reading this on, which are usually the same machine.

Frequently asked questions

Where does Zoom put my local recording?

By default, a Documents/Zoom folder with a subfolder per meeting. Inside you will usually find an MP4 of the meeting and an M4A of the audio on its own. Either works here; the M4A is smaller and is the faster choice if the video is large.

Do I need a paid Zoom plan for this?

No, and that is the point. Zoom's own cloud transcription is tied to cloud recording and to plan tier. Local recording is available much more widely, and once you have the file on your disk, transcribing it has nothing to do with Zoom at all. The same applies to Teams and Meet recordings you have downloaded.

Does a bot join my meeting?

Nothing joins anything. Meeting assistants work by adding a participant that sits in the call and records it, which is visible to everyone and often needs consent from the room. This works on a file after the fact, so there is no extra attendee and nobody has to be told a third party service is present.

Is the recording uploaded to transcribe it?

No. The file is read by your browser and the model runs on your own machine. For internal meetings this is usually the deciding factor, since a recording of a planning call is exactly the kind of thing a company would rather not have sitting in a vendor's storage bucket under whatever retention policy they happen to run.

Will it tell me who said what?

No. You get an accurate timestamped transcript without speaker labels. For a small call the turns are usually clear from context, and the timestamps let you jump back to the recording to check a specific moment. Automatic speaker attribution is a genuinely different feature and this does not do it.

Can it summarise the meeting or pull out action items?

Not on this page. This produces the transcript, accurately and privately. What you do with the text afterwards is up to you, and pasting it into whatever assistant you already use is a perfectly good next step now that the audio itself never had to be shared.