What this is
IMPDB is a database of podcast episodes about movies. You search for a film and get the episodes where that film is the subject, either for the whole episode or for a real segment of it. Episodes that only mention the film in passing are not included.
The site also does a few things beyond search. You can subscribe to a movie as a podcast feed, so new episodes about it show up in your podcast app. If you use Letterboxd, one feed can cover everything in your diary. You can follow shows, hide shows, and filter by language. All of that is described on the features page.
Why it exists
Podcast apps search show titles and descriptions. If you type in a movie, you get every show that has ever used the title, with no way to tell a two-hour discussion from a thirty-second aside. Apple and Spotify organize podcasts by show, not by subject, and neither one knows what an episode is about.
I wanted to finish a movie and then find the people who spent an hour on it. Nothing did that, so I built this.
How it works
A pipeline collects podcast feeds from Podcast Index, reads each episode's title and show notes, and tries to identify the movie. When the title names a film and a year, that is a match. When the title is a common word or phrase, or the film is only named in the description, a language model reads the text and answers one question: is this movie the main topic, a major segment, or a mention? Mentions are discarded and never stored.
The pipeline never listens to the audio. Everything it knows comes from the text a podcast publishes about itself. Shows that write real episode descriptions are covered well. Shows that title every episode "Episode 214" and leave the notes blank are mostly invisible to it.
What it gets wrong
Coverage is incomplete. There are feeds the pipeline has not found and episodes it has not classified yet. A model makes the topic call, and it is sometimes wrong in both directions: it misses episodes and it occasionally stores one that is not really about the film. Movies with common titles, remakes, and sequels are the hardest cases. If a show you know is missing, it is in the backlog or its feed does not say enough for the pipeline to work with.
Sources
Movie data comes from TMDB. This product uses the TMDB API but is not endorsed or certified by TMDB. Podcast data comes from Podcast Index and from the public RSS feeds themselves. Links open in Overcast, Apple Podcasts, Pocket Casts, and Spotify where an episode can be located there. Current counts are on the stats page.