# How to Download YouTube Transcripts from the Command Line (curl)

Source: https://transcript-yt.com/blog/youtube-transcript-curl-command-line · Updated 2026-11-09 · TranscriptYT

> Get YouTube transcripts from the terminal with curl and jq: plain text, SRT and VTT files, translations, and a Bash loop that downloads a whole list of videos.

You don't need Python or Node to get a transcript. One curl command against the TranscriptYT [YouTube Transcript API](https://transcript-yt.com/api) prints any public video's transcript, saves a subtitle file, or, in a short Bash loop, downloads a whole list of videos.

## Plain text in one command

Create a free key in the [dashboard](https://transcript-yt.com/dashboard/keys) (100 transcripts, no card) and export it. format=text returns the transcript without timestamps; leave format out for JSON.

```
export TRANSCRIPTYT_KEY="ts_live_..."

curl -s "https://transcript-yt.com/v1/transcript?url=5f3Pn-N6yLc&format=text" \
  -H "Authorization: Bearer $TRANSCRIPTYT_KEY"
```

## Subtitle files and translation

Use format=srt or vtt with -o to save a file, and translate_to to translate it into one of 150+ languages. --fail makes curl exit non-zero on an error instead of saving the error body.

```
curl -s --fail -o captions.es.srt \
  "https://transcript-yt.com/v1/transcript?url=5f3Pn-N6yLc&format=srt&translate_to=es" \
  -H "Authorization: Bearer $TRANSCRIPTYT_KEY"
```

## Pick fields from the JSON with jq

```
curl -s "https://transcript-yt.com/v1/transcript?url=5f3Pn-N6yLc" \
  -H "Authorization: Bearer $TRANSCRIPTYT_KEY" \
  | jq -r '.segments[] | "\(.start | floor)s  \(.text)"'
```

## Download a list of videos

Put one video ID or link per line in videos.txt. Failed requests aren't billed, so the loop just reports them and moves on.

```
mkdir -p transcripts
while read -r video; do
  id=$(basename "$video" | sed 's/.*v=//; s/[?&].*//')
  curl -s --fail -o "transcripts/$id.txt" \
    "https://transcript-yt.com/v1/transcript?url=$id&format=text" \
    -H "Authorization: Bearer $TRANSCRIPTYT_KEY" || echo "failed: $video"
done < videos.txt
```

## Whole playlists and channels

List a playlist's or channel's videos with /v1/playlist and feed the IDs into the loop above. The [channel and playlist guide](https://transcript-yt.com/blog/download-all-transcripts-youtube-channel-playlist) covers it end to end, and the [docs](https://transcript-yt.com/docs) list every parameter.

```
curl -s "https://transcript-yt.com/v1/playlist?url=https://www.youtube.com/@veritasium" \
  -H "Authorization: Bearer $TRANSCRIPTYT_KEY" \
  | jq -r '.videos[].videoId' > videos.txt
```

## FAQ

### How do I download a YouTube transcript with curl?

Run curl "https://transcript-yt.com/v1/transcript?url=VIDEO_ID&format=text" -H "Authorization: Bearer $TRANSCRIPTYT_KEY". Use format=srt or vtt for subtitle files.

### Can I download transcripts without yt-dlp?

Yes. yt-dlp downloads subtitle files from YouTube directly and often gets blocked on servers; the API returns the transcript over HTTPS from any machine, and creates one with AI when a video has no captions.

### How do I download transcripts for a whole playlist?

Call /v1/playlist to list the video IDs, then loop over them with the command in this guide, or use the batch endpoint to fetch many in one request.