Podcast search with a number that says who is big

No. A live episode lookup returned published as the string "2026-08-26T02:45:00-07:00", ISO-8601 with a UTC offset. A second episode on an unrelated show returned "2025-06-26T00:02:00-07:00", same shape. Parse it as a datetime with offset; do not divide it by 1000 or hand it to a unix-epoch constructor.

Because that is what Listen Notes publishes. It is a percentile band, not an ordinal position, so there is no rank 1, 2, 3 to give you. Sort by listen_score, a 0-100 integer, when you need a numeric ordering, and expect ties inside a band: three of five measured US rows shared "TOP 0.5%" at listen scores of 52, 52 and 54.

No. Listen Notes only assigns a Listen Score to part of its index, and non-US lists are thin. Measured: every row of best_podcasts for region us carried both fields, and every row for region se carried null for both. Null means unscored, not zero, so drop those rows from a ranking rather than treating them as the worst entries.

No. Three ids coexist and each action takes exactly one of them. The podcast action takes the 11-character id (xp7nhsmSkX2); the 32-hex uuid on the same object is a different identifier and is not a lookup key. The episode action takes a 32-hex episode id, which looks like a uuid but belongs to an episode, not a show. When in doubt, pass the full listennotes.com URL instead, which both actions accept.

You get a clean NOT_FOUND rather than an empty list. genre_id 999999 returned error code NOT_FOUND naming the 404 URL it tried. That distinction matters, because an empty 200 would look like "this genre has no podcasts", which is a different claim. Call genres first: it returned 159 entries with ids from 68 to 266, and the ids are not contiguous, so you cannot iterate a range.

Listen Notes prefixes the episode description with the runtime, so the text arrives as "00:50:06 - This summer, the Democratic Socialists...". It is upstream formatting rather than a parsing artifact, and the same prefix appeared on episodes from two unrelated shows. Strip the leading "HH:MM:SS - " before you index the text or feed it to a model; the same number is already there as audio_length_sec.

Because Listen Notes puts the raw RSS URL behind its own login for most shows, and the response says so: rss was null on both podcasts measured, with rss_login_required true in the same object, and itunes_url was null alongside it. What you do get is extra_urls, which carried working Spotify, Amazon and Google links plus Twitter and Facebook handles for The Daily.

Ten, regardless of how many the show has. The Daily reported total_episodes 2961 and returned 10 rows in latest_episodes; The Startup Solution reported 41 and also returned 10. Those rows carry episode_id, title, audio, duration and duration_sec, which is enough to hand each one to the episode action for the full record including description and publish date.

Listen Notes API is a ReefAPI endpoint group for listen notes It returns live JSON through POST requests under /listennotes/v1.

Yes. ReefAPI starts with 1,000 free credits, no card required. Listen Notes calls use the same shared credit balance as every other ReefAPI engine.

No login to Listen Notes is needed for the API response. You call ReefAPI with your x-api-key header, and the playground can run live examples before you create a production key.

The page example is captured from a live search call, and production requests fetch live data through ReefAPI rather than a static sample.

Listen Notes actions currently cost 1 credit per successful call. Failed or blocked calls are free, and all APIs draw from one credit pool.

Yes. Connect ReefAPI once through MCP and your assistant can call listennotes actions with the same key, credit pool and JSON envelope used by normal REST requests.