Podcast transcription, one API call.

Podcast Transcript API: Spotify, Apple and RSS episodes to text

100 free credits · no card

Spotify, Apple Podcasts or a raw RSS enclosure. We resolve the link to the real episode audio and transcribe it, however long it runs. Per-segment start times mean you can quote the exact minute a claim was made. One credit per episode, whether it is twenty minutes or three hours.

episode · 1:42:11200 OK
12:04The mistake was assuming the first version had to be the one we shipped forever.
12:11We rewrote the pricing page four times before anything moved at all.
12:19And the version that worked was the one that said the least, which surprised everyone.
12:27People were not confused about the features, they were unsure whether it was for them.
12:35So we cut the feature list in half and put the use case at the top instead.
12:44Signups went up without a single change to the product underneath it.
transcribed · en · 2,184 segments1 credit

Podcasts are where per-minute pricing hurts.

A three-hour episode costs one credit. Transcription services that bill per audio minute charge for every one of those 180 minutes.

With a transcription service

You are billed per audio minute, so the long episodes you most want to search are the ones you can least afford to process. You also have to find and download the audio yourself, which means resolving a Spotify or Apple link to a real file before you start.

With TranscriptFetch

Send the episode link. We resolve Spotify, Apple Podcasts and RSS to the underlying audio and transcribe it, at one credit regardless of runtime. Failed requests cost nothing, and a fetch that cannot be resolved tells you why.

Who it is for, and how they use it

Three ways in, depending on how you work. Same endpoint, same credit, same response shape underneath all of them.

DevelopersOne endpoint

Engineers building a product

You need episode text in a pipeline and you do not want to own the resolve-then-transcribe plumbing.

  1. 01Create a key in the dashboard, send it as a Bearer token.
  2. 02POST the episode link and get a job back to poll.
  3. 03Store the segments next to your other sources and move on.
Read the API reference
Research and archivesRun at volume

Analysts and researchers

You have a back catalogue nobody has ever searched, because listening to it end to end is impossible.

  1. 01Point at the feed and collect the episode links.
  2. 02Send up to 50 in one batch call.
  3. 03Search the text and quote the minute something was said.
See batch fetching
AI and agentsNo code

Teams building with assistants

Your agent needs to answer from an episode and cite the minute, not paraphrase a summary of it.

  1. 01Add the MCP server once, no deploy and no glue code.
  2. 02Ask in plain language, the assistant calls the tool itself.
  3. 03Answers come back with the timecode attached.
Set up MCP

One request, whichever way you build.

Write code if you want to. If you would rather not, the same fetch runs from a workflow tool or straight inside an assistant.

REST

Call the endpoint

One POST with the episode link. Long audio returns a job to poll.

Read the API reference
NO CODE

Drop it in a workflow

An official n8n node. Transcribe new episodes as the feed updates.

See the n8n node
MCP

Let an assistant do it

Connect once and Claude or ChatGPT reads the episode mid-chat.

Set up MCP
curl -X POST https://transcriptfetch.com/api/v1/transcripts/video \
  -H "Authorization: Bearer $TRANSCRIPTFETCH_KEY" \
  -d '{"video": "https://open.spotify.com/episode/4rOoJ6Egrf8K2IrywzwOMk"}'

# 202 + job_id, then poll:
# GET /api/v1/transcripts/jobs/{jobId}
No code required

Let the assistant read the episode itself.

Add the MCP server once and Claude, ChatGPT or Cursor can pull an episode transcript mid-conversation. Same key, same credit, no glue code and nothing to deploy.

https://transcriptfetch.com/mcp

ClaudeTool: transcriptfetch
What did they say about pricing in this episode?
CALLING get_transcript(video="open.spotify.com/episode/4rOo…")
At 12:35 he says cutting the feature list in half raised signups without touching the product.

Questions people ask first

Podcast Transcript API vs a per-minute transcription service

TranscriptFetch
A per-minute transcription service
Pricing
One credit per episode, whatever the runtime
Billed per audio minute; a three-hour episode is 180 of them
Finding the audio
Spotify, Apple and RSS links resolved for you
You locate and download the file first
Show metadata
The show and episode come back named
A filename to parse
Timestamps
Per-segment start and duration
Varies by service
Long episodes
Up to four hours, chunked automatically
Upload limits; split the file yourself
Batch
Up to 50 episodes per call
One upload at a time
Whole back catalogues
Resolving one episode indexes the entire feed
Manual, file by file
Failed fetches
Free, billed on delivery only
Charged for the minutes processed
MCP server
Yes, for Claude, Cursor and others
No

Neither Spotify nor Apple publishes a public transcript API. RSS is the source of record and both platforms are directories that ingest it, which is why all three link types resolve to the same episode audio here.

Why caption-only tools return nothing here

TranscriptFetch
A caption-only fetch
Podcast episodes
Audio transcribed with AI, every time
No caption track exists to read
Spotify or Apple link
Resolved to the real episode audio for you
Not resolvable
Changes needed in your code
None, same video endpoint
A whole transcription pipeline
Extra cost
Charged only when the transcript is delivered
Unavailable at any price

What the endpoint returns

The full episode text, however long it runs
Per-segment timestamps: start and duration
A podcast block naming the show and the episode
Spotify, Apple Podcasts and raw RSS links accepted
Episodes up to four hours, chunked automatically
Batch endpoint for up to 50 episodes per call
A job to poll, or a webhook when the transcript is ready
An MCP server, plus Python and JavaScript SDKs

Guides from the blog

Start with 100 free credits

Paste an episode link or copy the curl, whichever suits you. No card, no sales call.