Burn a Zoom Transcript Into a Video With FFmpeg

Zoom already puts the transcript inside your recording — it just hides it. Zoom's audio transcription page states it in one line: "The transcript is automatically embedded within the audio and video file, but is hidden by default."

What form that embedded transcript takes inside a file you have downloaded is something Zoom's official documentation doesn't state. Burning the text in sidesteps the question: the words are drawn into the frames themselves, so there is nothing left to toggle. That step happens outside Zoom, and one ffmpeg command covers it.

The command

ffmpeg -i video98796876.mp4 -vf "subtitles=transcript.vtt" -c:a copy burned.mp4

That odd input name is not a placeholder I invented. Zoom's file format reference says the video file "is named video[random number].mp4 by default," and gives video98796876.mp4 as its own example. If you went looking for a file named after the meeting topic and found nothing, that is why.

What each flag does

• -i video98796876.mp4 — the recording you downloaded from the Zoom web portal. Point it at the download, not at a re-share or a screen capture of playback.

• -vf "subtitles=transcript.vtt" — the subtitles video filter, and the only part of this command doing real work. Because it is a video filter, the picture is decoded, the text is drawn on top of every frame, and the result is encoded again. That is precisely what "burned in" means: hard subtitles rendered into the pixels. Turning off a caption track afterwards does nothing, because there is no caption track — there is just video that has words in it.

  The filter reads the subtitle file directly. Clapr's own burn tool builds the same subtitles= filter and accepts .srt, .ass, and .vtt, so a transcript straight out of Zoom needs no conversion step first.

• -c:a copy — leave the meeting audio alone. FFmpeg's documentation describes stream copy as copying an input stream's packets "without decoding, filtering, or encoding them," and spells out the consequence: "Since there is no decoding or encoding, it is very fast and there is no quality loss." Only the picture is being rebuilt here, so there is no reason to touch the audio.

• burned.mp4 — a new file, deliberately. Never write over the Zoom download. If the source was a computer recording, Zoom is blunt about what happens if you lose it: "Local recordings exist only on the computer where they were created. Zoom does not retain a copy and cannot recover a lost local recording."

Which of Zoom's VTT files you actually want

Zoom can hand you two different VTT files, and they are not interchangeable. The format reference lists them separately:

"CC.VTT: Text file that contains closed captioning for the cloud recording."

"VTT: Text file that contains the audio transcription for the cloud recording."

Zoom's developer documentation draws the same line, with TRANSCRIPT described as "Transcription file of the recording in VTT format" and CC as "File containing closed captions of the recording in VTT file format." They also come from two different switches in the cloud recording settings — "Create audio transcript" for the first, "Save closed caption as a VTT file" for the second.

Two things follow that catch people out:

1. Transcripts are cloud-only. Zoom states that "Audio transcription only applies to cloud recordings, not computer recordings that are saved to your device," and lists audio transcription among the features computer recordings do not support. If you recorded to your laptop, there is no VTT to burn and never will be.
2. The transcript lands after the video. Zoom notes that "Once the cloud recording becomes available, the audio transcript may take additional time to process." A recording that is ready with no VTT next to it is usually not broken — it is early.

Fix the text before you burn it

Burned text is permanent, so the editing pass matters more here than it would for a sidecar file. Zoom flags exactly what will be missing:

"The transcript is divided into sections, each with a timestamp showing how far into the recording that portion was made. You can edit the text to more accurately capture the words or to add capitalization and punctuation, which are not included in the transcript."

Capitalization and punctuation are not included. Read that again before you commit an hour of meeting to pixels. Fix it in the web portal, or download and edit locally — Zoom notes that "Audio transcripts are saved in VTT format. If you download the VTT file, you can open it with a text editor or word processing application."

On language, Zoom's own pages disagree, and it is worth knowing both. The page on enabling transcription states flatly that "Audio transcription only supports English." The page on using transcription says transcripts "are supported in multiple languages," lists nineteen of them, and documents changing a recording's language so that "Zoom will reanalyze and transcribe the audio in the newly selected language." Both are live official pages. FFmpeg itself is indifferent — it renders whatever text the file contains.

Variations

Make the text bigger or reposition it with force_style, which overrides the subtitle styling on the way in:

ffmpeg -i video98796876.mp4 \
  -vf "subtitles=transcript.vtt:force_style='Fontsize=32,Outline=3,MarginV=56'" \
  -c:a copy burned.mp4

Those are the same three values Clapr's burn tool uses for its "large" style — font size, outline width, and vertical margin.

If your pipeline wants SRT instead, the filter takes either format, so convert only when something downstream demands it. Use VTT to SRT — and know that the conversion is lossy in a specific way: cue identifiers get renumbered and cue positioning settings are dropped, because SRT has nowhere to put them.

If the burned file is going straight to the web, the moov atom placement is a separate concern from subtitles entirely; see ffmpeg and the faststart flag.

If you would rather not install ffmpeg

For one recording, a browser tool is faster than a toolchain. This one runs the same subtitles= filter compiled to WebAssembly, inside the tab — your recording is read from disk, processed locally, and handed back, and nothing is uploaded to a server. For a meeting recording that is usually the deciding factor.

If you also need the container changed rather than just captions added, converting a Zoom recording to MP4 covers what Zoom does and does not tell you about its output files.

Troubleshooting

There is no download button in the web portal. Zoom's answer is unexpectedly literal: "Make sure your browser window is set to a minimum width of 1024 pixels. Maximize your browser window and set your monitor resolution to at least 720p (1280 by 720 pixels)."

The download is blocked entirely. If you are not the host, Zoom says "The recording host may have restricted access to download the recording, meaning you can't download the recording even if you use a tool to search the URL for a download. Contact the host to gain access."

The captions drift out of sync. Check whether the recording was trimmed after you pulled the VTT. Zoom warns that trimming "affects all associated audio and transcript files, overwriting the original" — so a transcript downloaded before a trim describes a video that no longer exists. Grab both files after any trim.

Do not assume what is inside the MP4. Zoom describes it only as an "Audio/Video file" and never names the codecs; for anything that depends on knowing them, check the official documentation rather than guessing.

FAQ

Can Zoom burn the transcript into the video for me? Zoom documents displaying the transcript "within the video itself, similar to a closed caption display" during playback, and a sharing option where "Viewers can see transcript." What it does not document is what form the embedded transcript takes inside a file you have downloaded. If you need the text visible everywhere, render it yourself.

Where is the transcript file in my Zoom recording? In the recording's file list, as a separate VTT. Watch for two VTT files: the plain VTT is the audio transcription, and CC.VTT is closed captioning. They come from different settings and are not the same text.

Why does my recording have no transcript at all? Most often because it was a computer recording — Zoom states audio transcription applies only to cloud recordings. Otherwise, the transcript may still be processing, since Zoom says it can take additional time after the recording itself becomes available.

Does burning captions reduce the video quality? The picture is re-encoded, because drawing text onto frames requires decoding and encoding them. The audio is not: -c:a copy passes it through, and FFmpeg's documentation states there is no quality loss in a stream copy.

Can I un-burn the captions later? No. That is the trade you are making. Keep the original Zoom download and the original VTT, and treat the burned file as a derivative you can always regenerate.

Zoom behaviour cited from Zoom's official support articles on recording file formats, audio transcription for cloud recordings, cloud recording settings, and trimming cloud recordings, checked 16 September 2026. FFmpeg stream copy behaviour cited from the official FFmpeg documentation. See all Clapr tools or the rest of the ffmpeg guides.