Skip to content

Docs · MCP server

TranscriptLab MCP server

Last updated October 20, 2026

TranscriptLab is a remote Model Context Protocol (MCP) server. It gives Claude and other MCP clients tools to fetch YouTube transcripts, transcribe audio and video files, search the results, and export them. For the REST API, see the API reference.

Server details

  • URL: https://api.transcriptlab.dev/mcp
  • Transport: streamable HTTP. No local install and no API key required to connect.
  • Authentication: OAuth 2.0 with dynamic client registration (RFC 7591).
  • Works with: Claude (claude.ai, Claude Code, Cowork) and any MCP client that supports remote servers.

Connect

claude.ai. Open Settings, then Connectors, choose Add custom connector, and paste https://api.transcriptlab.dev/mcp. You can also find TranscriptLab in the Claude connectors directory.

Claude Code.

claude mcp add --transport http transcriptlab https://api.transcriptlab.dev/mcp

Other MCP clients. Point the client at the URL above. It discovers the sign-in endpoints on its own and registers itself, so no client ID is needed.

Sign in

Connecting and listing the tools needs no sign-in. The first tool call opens a TranscriptLab consent page where you either sign in with your TranscriptLab account or paste a TranscriptLab API key. Create an account at transcriptlab.dev.

  • Access tokens are scoped transcriptlab:read and transcriptlab:write and last one hour. They refresh automatically.
  • A connection made by pasting an API key can use every tool except those that need a signed-in session: download links and large-file uploads.
  • Remove access any time by deleting the connector in your client, or by revoking the session key in your TranscriptLab dashboard.

Tools

Every tool declares a title and read-only or write annotations. Tools that fetch or transcribe are marked as writes because they create a transcript in your library.

ToolWhat it doesCost
transcriptlab_get_transcriptGet the transcript of a YouTube video from its captions, optionally in another language.First delivery of a variant can use a credit
transcriptlab_retrieve_transcriptRead a YouTube transcript that is already published for your account, without starting a new fetch.Free for variants you already have
transcriptlab_discover_captionsList the caption languages and track types (manual or generated) a video offers.Free, subject to a monthly limit
transcriptlab_get_job_statusCheck whether a transcript job has finished.Free
transcriptlab_search_transcriptSearch, quote or page through a transcript you already have (keyword, regex or sequential chunks).Free
transcriptlab_list_transcriptsList the transcripts in your library, newest first.Free
transcriptlab_transcribe_fileTranscribe an audio or video file attached to the conversation with speech-to-text.Credits per audio minute
transcriptlab_transcribe_media_urlTranscribe a public https link to an audio or video file.Credits per audio minute
transcriptlab_get_speech_transcriptRead the speech-to-text result for a file or media link you submitted.Free
transcriptlab_export_linkCreate a 10-minute download link (TXT, SRT, VTT, Markdown or JSON) for a transcript you have.Free; needs a signed-in connection
transcriptlab_get_usageShow your plan, remaining credits and usage this period.Free
transcriptlab_openOpen the interactive TranscriptLab view: paste a link, upload a file, browse your library.Free
transcriptlab_open_audioEntry point a host uses when you open an audio or video file with the connector. You do not call it yourself.Credits per audio minute

Examples

  • “Get the transcript of https://www.youtube.com/watch?v=dQw4w9WgXcQ and summarize it.”
  • “Which caption languages does that video have?”
  • “Find where they mention pricing in that transcript.”
  • “Transcribe this attached interview recording and label the speakers.”
  • “Transcribe https://example.com/interview.mp3.”
  • “Give me that transcript as an SRT file.”
  • “Show my earlier transcripts” and “How many credits do I have left?”

Credits and limits

  • Fetching a YouTube transcript uses a credit the first time you request that variant. Reopening, searching and exporting a transcript you already have is free.
  • Speech-to-text is metered per minute of audio. The standard model uses 1 credit per minute; the premium model, which can label speakers, uses 10.
  • Caption discovery is free but limited each month by plan.
  • Files attached in chat can be up to 300 MB. Larger files, up to 2 GB, go through the upload view. Media links must be public https URLs.
  • When you are out of credits or at a limit, the tool says so and stops. It does not retry.
  • Download links expire after 10 minutes.

What gets sent and stored

The server receives the YouTube links, files and media URLs you ask your client to process, and the key you paste on the consent page if you choose to. Delivered transcripts stay in your TranscriptLab library until you delete them; you can delete individual transcripts or your whole account. A YouTube video without captions is not transcribed automatically: the client says so and offers speech-to-text if you can provide the file or a direct link.

See the privacy policy and terms.

Troubleshooting

  • “Connect your TranscriptLab account to use this tool.” Finish the sign-in prompt, or reconnect the connector once and try again.
  • A transcript is “still processing”. Long videos can take a minute. The client checks the job and reads the result when it is ready.
  • No captions found. The video has no caption track. Use speech-to-text with an attached file or a direct media link.
  • Download link will not open. Some hosts block automatic downloads. Use the link shown on screen within 10 minutes, or the Copy button in the view.
  • Out of credits. Check your balance on the dashboard or ask for your usage, then top up on your plan.

Support

Email hello@transcriptlab.dev and we will reply, usually within one business day. See also the support page.