In ChatGPT or Claude
Connect TranscriptFetch as an MCP server and ask your assistant for a transcript, a channel's uploads, or a search. It fetches the data and works with it in the same conversation.
Connection guidesCapabilities
What TranscriptFetch can do
One service for turning video into text you can work with. Here is every capability, where each one is available, and the limits worth knowing before you build on it.
What you can do
- POST/api/v2/transcripts/video
Get a transcript with timestamps
Send a video link and get the spoken words back as timestamped segments and plain text. Captions are used when the platform has them, so the result is exact and instant.
- GET/api/v2/transcripts/jobs/{jobId}
Transcribe when there are no captions
Most TikToks and Instagram videos have no caption track. When one is missing, the audio is transcribed with AI and the result comes back in the same shape, with timestamps.
- POST/api/v2/transcripts/channel
List a channel or profile
Get the latest uploads from a YouTube channel, a TikTok profile, or an Instagram account as a list you can transcribe, filter, or check again on a schedule.
- POST/api/v2/transcripts/playlist
List a playlist
Turn a YouTube or TikTok playlist into a list of videos, then fetch transcripts for the ones you want.
- POST/api/v2/transcripts/search
Search for videos
Search by keyword on YouTube, TikTok, or Instagram and get results you can transcribe straight away.
- POST/api/v2/transcripts/batch
Fetch transcripts in bulk
Send a list of links in one call and get transcripts back for all of them. The batch size depends on your plan.
Where it works
Which capabilities are available on each source. Accepted link formats for every platform are listed in the docs.
| Transcript | Channel or profile | Playlist | Search | |
|---|---|---|---|---|
| YouTube | available | available | available | available |
| TikTok | available | available | available | available |
| available | available | not available | available |
How to use it
Every capability is available from all three, so pick whichever fits the way you work.
From your code
Call the REST API from any language, or use the Python and Node.js SDKs. Every capability above is one endpoint.
API referenceIn the Playground
Try any request in the browser with your own key before you write code, and copy the working call out.
Open the PlaygroundGood to know
- Credits
- One credit per successful caption fetch or listing. AI transcription is billed per started minute of audio, and only when a transcript is delivered. Failed requests are free.
- Words on screen are not speech
- Text a creator typed onto the video in an editor is pixels, not audio, so it does not appear in the transcript.
- No speech, no transcript
- A clip with only music, or a silent slideshow, returns no transcript rather than an invented one.
- Public content only
- Private, removed, or login-gated videos cannot be fetched. Live streams can be transcribed once the broadcast has ended.
Try it on a video
Sign in, paste a link into the Playground, and read the transcript in seconds.