Tracking new episodes as they’re ingested? Don’t poll this list with
since — use the episode feed, a resumable poll that returns each new episode exactly once, or the real-time episode stream (Enterprise).Available to MCP agents as
particle_podcast_list_episodes and (per-episode) particle_podcast_get_episode.List episodes
Response (truncated)
has_transcript, segment_count, clip_count, and entity_count fields tell you what enrichment data is available before you fetch any sub-resource. transcript_status says where the episode stands on the way to a transcript; see transcript status and the back catalogue.
Filter parameters
Holding an Apple Podcasts episode ID (the
?i= value in an Apple
Podcasts URL) or a Spotify or YouTube episode link instead? That resolves
through GET /v1/podcasts/episodes/lookup,
not a filter here.
Combine freely — for long-form transcribed Sam Altman episodes from this month, for example:
Get a single episode
Which audio URL to use
Each episode carries two audio URLs:urlis the audio file as the podcast’s feed publishes it. It often passes through a measurement or ad-insertion service such as Podtrac, which counts the download and can insert different ads each time. Use it to play the whole episode to a listener, so the publisher’s downloads and ad impressions are counted.audio_urlis our CDN copy, and every timestamp in our responses (transcript lines, segments, clips, mentions) is measured against it. Use it when no one will listen, and whenever you play or cut audio from a specific point: the ads inserted intourlcan vary in length, so its timing won’t match ours.
Episodes for a specific podcast
A convenience shape oflist-episodes filtered by podcast:
Transcript status and the back catalogue
Every episode carriestranscript_status:
When a show joins the catalog we ingest and transcribe its newest episodes, and every episode it publishes after that. Its older episodes — up to the 1,000 most recent PodcastIndex lists, plus anything else its own feed still carries — are its back catalogue: listed with their feed metadata, but with no transcript, speakers, entities, topics, segments or clips.
The episode lists leave the back catalogue out unless you ask for it with
transcript_status, so an integration that pages through episodes to fetch transcripts sees exactly what it always has. To see a show’s whole history, pass podcast_id and one of:
any: every episode we know of, newest first, transcribed or not.untranscribed: episodes without a transcript, whether processing, unavailable, or in the back catalogue.requestable: back-catalogue episodes only, including those alreadyqueued.
duration_seconds is the length the publisher declares, and it has no slug or audio_url until it is transcribed. Until it is transcribed, which you can request, its transcript, segments, speakers, entities, topics, ads, clips and related episodes answer 409 transcript_not_available rather than 404, so an episode that exists but is untranscribed is never mistaken for one that does not exist. An unavailable one answers 422 episode_audio_unavailable. GET /v1/podcasts/episodes/{id} returns it by id, and episode lookup resolves its Apple Podcasts ID, feed guid or PodcastIndex ID. Filters that select transcript-derived content — entity_id, person_id, company_id, fully_ingested — match no back-catalogue episodes.
To see how much of a show’s history exists and how much is transcribed without paging its episodes, request GET /v1/podcasts/{id}?include=coverage. The coverage section counts discovered episodes by transcript_status and by year, as of counted_at (the counts refresh hourly; each episode’s own transcript_status is always current), and says whether the show’s back catalogue has been imported and whether it is complete (back_catalogue.complete is false when PodcastIndex listed its maximum of 1,000 episodes, so older ones may exist that neither source lists):
episode_count on the podcast counts only the episodes we have ingested; use coverage.discovered_episode_count for the show’s whole known history, and remember that counts derived from transcripts, such as sponsors or guest appearances, cover only the transcribed episodes.
Sub-resources
Every episode exposes its enrichment data as sub-resources:Speakers
Response (truncated)
HOST,
GUEST, PANELIST and CORRESPONDENT. Use role to widen or narrow the set;
values are case-insensitive, and roles always come back in uppercase.
Response (truncated)
Speaker roles
Every speaker carries exactly one of these roles. The four in the default set are marked; the rest require naming them inrole.
This list is exhaustive and stable, across the whole archive — you can branch
on
role without a fallback case.entity_slug (when present) bridges to the knowledge graph.
Entities
Returns people, organizations, places, products, and concepts mentioned in the episode, ranked by salience — a relative relevance score for that episode where higher values indicate the entity is more central to the discussion.Response (truncated)
Topics
Related
- Transcripts — dialogue, words, mentions, SRT
- Segments & clips — structural breakdown and highlights
- Knowledge graph → entities — follow speakers and topics across the catalog