Skip to main content
Audivo turns a public podcast episode into a clean, timed, provenance-carrying transcript over an HTTP API, and exposes the same operations to AI agents as an MCP server.

​
How a request flows

  1. Send the episode to POST /v1/transcripts: an Apple Podcasts link, a feed URL and GUID, an episode id, or an upload of your own audio.
  2. Get it back. An episode already transcribed comes back at once; anything else is a job that finishes in a minute or two.
  3. Read the transcript as JSON, plain text, SRT, VTT or Markdown.
That is the whole path for one episode. For many at once — a chart, a back catalogue — price the selection first and confirm it as one group. Credits are the unit of spend: one credit per audio minute, rounded up per episode. A job holds a ceiling of its estimate plus 25% and settles at the measured audio, never above. max_credits on a call refuses anything that could cost more than you allow.

Quickstart

One episode, one call, with curl.

Authentication

API keys, the one header every call sends, and what a 401 looks like.

MCP server

Connect Claude, ChatGPT, Codex, Cursor and other clients.

Errors

Every error code, what it means, and whether to retry.

Upload your own audio

Transcribe a recording, or a show with no public feed whose audio you obtained yourself.

Price a selection

Many episodes at once: quote, confirm, and read a job group.