Every show, every episode, everyone who appeared on them.
pods.wiki reads the podcast circuit first-party from Apple, RSS, YouTube, Spotify and the Podcast Index, deduplicates it, and stores every claim with the source it came from. Then it makes the whole thing searchable by meaning rather than by keyword.
What you get
Identity that holds
The same person is never in the table twice. Strong external keys resolve a known handle deterministically, and merges are non-destructive, so an id you were handed a year ago still resolves today.
Provenance on every field
Each claim records where it came from. A reference work that cannot say why it believes something is not much use, so this one always can.
Search by meaning
Transcripts, themes and show notes are embedded, so you can ask for the idea rather than guessing the phrase somebody used.
Reading a record is one credit. Asking a question that spends a model call is two hundred. The price list is generated from the running API, so it can never disagree with what you are charged.
| 0 | Free. Liveness and the catalogue itself. | 2 routes |
| 1 | One record, read by id. | 8 routes |
| 3 | A filtered, paginated listing. | 14 routes |
| 5 | One subject, joined across the corpus. | 18 routes |
| 10 | An aggregate over the whole corpus. | 9 routes |
| 25 | A heavy aggregate. Slow to build, and priced to say so. | 6 routes |
| 50 | Spends an embedding on your text. | 1 route |
| 200 | Spends a model call. | 1 route |
| 500 | Spends research budget on a live lookup. | 2 routes |