Transcription for your AI assistant

convert.express runs an MCP server, so Claude, ChatGPT, Cursor and other assistants can transcribe audio and video for you. Give your assistant a link or a file and ask for a transcript, subtitles, a summary or a translation. No account and no API key.

https://convert.express/mcp

Connect it

Claude Code

claude mcp add --transport http convert-express https://convert.express/mcp

Claude (web and desktop app)

Open Settings → Connectors, choose Add custom connector, name it convert.express and paste the URL above. Leave authentication empty.

ChatGPT

Custom MCP connectors are behind developer mode. Turn it on in Settings → Apps & Connectors → Advanced settings, then create a connector with the URL above and no authentication. OpenAI moves these menus from time to time; if the names differ, look for a custom connector or MCP server option.

Cursor

Add this to ~/.cursor/mcp.json (all projects) or .cursor/mcp.json (one project):

{
  "mcpServers": {
    "convert-express": { "url": "https://convert.express/mcp" }
  }
}

VS Code

Add this to .vscode/mcp.json:

{
  "servers": {
    "convert-express": { "type": "http", "url": "https://convert.express/mcp" }
  }
}

Anything else

Any MCP client that speaks Streamable HTTP works. The server is stateless and needs no headers beyond the MCP defaults.

Try it

Once connected, ask in plain words. For example:

“Transcribe this YouTube video and give me the action items: <link>”
“Make German subtitles for ~/Videos/talk.mp4.”
“Transcribe interview.m4a with speaker labels and send me a share link.”

Uploading a file from your computer needs an assistant that can run commands or make HTTP requests, like Claude Code or Cursor. Chat apps work best with links.

What the assistant can do

transcribe_url: Transcribe a public link: a direct audio/video file or a YouTube video.
create_upload, start_transcription: Upload a file from your machine, then start the job.
get_job: Status and results: text, SRT or VTT subtitles, timestamps with speakers, download links, and one ZIP with everything.
request_summary: An AI summary or a list of action items.
translate: Translate the transcript, subtitles (timecodes kept), summary or action items into any of 103 languages.
create_share_link: A web page with the results that you can open or send to someone.
list_jobs, delete_job: See your jobs, or delete one now instead of waiting 24 hours.
estimate_cost, get_service_info: Prices, models, languages and limits.
send_feedback: Report a bug or ask for a feature. A person reads it.

Models, languages and formats are the same as on the website: OpenAI Whisper, MAI-Transcribe 2 and Gemini 3.5 Transcribe, files up to 2 GB. See models and languages.

Free use and paying

Assistants get some free use every day. Past that, your assistant gets a preview (the first 100 words), the price and a payment link, and it will show you the link. Open it, sign in and pay: €0.02 per minute, €0.50 minimum, no subscription. Your own free daily minutes may lower the price. Then the assistant fetches the full transcript.

After that first sign-in, the assistant is linked to your account. Its next jobs go straight into your dashboard and use your free minutes. You can unlink it on the dashboard under Connected AI assistants. Full pricing: pay as you go.

Privacy

Files and results are deleted 24 hours after the last activity, or right away with delete_job.
Your files are never used to train AI models.
We log each call from an assistant (IP address, the client name it reports, which tool, when) for 90 days to stop abuse. Transcript text is never logged.

Details in the privacy policy.

Problems or ideas

Ask your assistant to use send_feedback. A person reads every message.