Transcription API for podcasts, YouTube and TikTok.
POST a link, take a webhook, pull JSON, SRT, VTT, DOCX or PDF — plus chapters, show notes and translations from the same transcript. Over HTTP, or over MCP.
Poll that id or register a webhook. ?include=text for the body, ?format=srt for a file. Full reference in the docs.
MCP
Connect an agent
One command. No SDK, no config file.
Terminal
# 1 · Connect the server
claude mcp add --transport http podcaststotext https://podcaststotext.com/api/mcp \
--header "Authorization: Bearer ptt_your_key_here"
# 2 · Or bridge it locally, for stdio-only clients
PODCASTSTOTEXT_API_KEY=ptt_your_key npx -y podcaststotext-mcp
# 3 · Then ask, in the assistant:
# "Transcribe the latest Hard Fork episode and give me chapters"
# Optional · the calling order, not just the tool list
npx skills add marstyl99/podcaststotext-agent
Yes. The API and the MCP server are on Creator, Pro, Studio and Podmaxxing The free tier is for trying the studio, not the API. The entry plan raises the monthly allowance and the rate limits and unlocks the AI tools; extra API minutes are sold in packs, spendable while your plan includes API access, and paused rather than lost if you cancel.
Yes. YouTube is read from the video’s existing caption track, so it completes in about a second and costs no credits, but a video with captions disabled cannot be transcribed this way. TikTok is downloaded and transcribed normally.
Yes, at /api/v1/openapi.json. It is generated by a route rather than kept as a file, so the server URL follows the deployment, and a test checks it against the actual route tree.
Get a key and make the first call
The API and the MCP server are on Creator, Pro, Studio and Podmaxxing. Keys are scoped and revocable, and one key drives REST and MCP alike.