Find the Episode You Half-Remember

No. This tool searches only the text a show publishes in its RSS feed: the episode <title> and the episode description (the <description> element, or <itunes:summary> where a host uses it instead). It does not read, generate, or search transcripts of the audio. Transcript-based search is a separate paid feature inside The Podcast App and is not what this page does. If a topic is discussed in an episode but never written into the title or show notes, this tool cannot find it.

Up to the 200 newest dated items in the feed. The tool requests the feed with limit=200, which is the maximum the endpoint accepts, and searches every episode it receives. If the feed contains more items than that, the results header tells you exactly how many of how many are being searched, so you always know the scope of the answer.

Yes. If what you type is not an http:// or https:// URL, it is treated as a show name and resolved against the Apple Podcasts catalogue to find that show's public feed URL, then the feed is searched. Name lookup can fail or match the wrong show when several podcasts share a title, so pasting the exact feed URL is always the more precise option.

Because that item's <pubDate> is missing, empty, or not in the RFC 2822 format that RSS requires (for example Tue, 12 Aug 2025 09:00:00 +0000). Malformed dates are a common export bug in podcast CMS platforms. Undated items are still searched, but they are listed after the dated ones instead of being sorted into the timeline, because there is no trustworthy date to sort them by.

Duration comes from the <itunes:duration> element. Some feeds omit it entirely, and some publish it in a form that cannot be read as a number of seconds. When that happens the episode is shown without a duration rather than with a guessed one. The audio file itself is not downloaded to measure its length, so no duration can be inferred.

Only what the feed itself publishes. An RSS feed is not an archive guarantee: many hosting platforms publish a rolling window of recent episodes rather than every episode ever released, and some shows deliberately remove old items. If a show's oldest episodes are not in its live feed, no feed-reading tool can search them. The tool reports the episode count it actually received, so you can compare it with the show's published total.

Matching is case-insensitive substring matching, and every whitespace-separated term must appear somewhere in the title or the description for an episode to match — an AND, not an OR. There is no stemming, no fuzzy matching, and no synonym expansion, so talks will not match talking. Each matched term is highlighted in the snippet, and every episode also shows a numeric match count, so a result is never signalled by colour alone.

The feed you request is fetched server-side and its parsed result is cached for one hour, keyed by the feed URL and the requested episode limit, so a repeated search does not re-download the same feed. Responses carry an X-Cache header of HIT or MISS so you can tell whether you received a cached copy. Filtering by your search term happens in your browser against the already-loaded episodes, so the words you type are not sent to the server.

Yes. The feed and the query are written into the page URL as ?feed= and ?q= parameters as you search, so the address bar always reflects the current search. Copy that link, or use the Copy results link button, and whoever opens it lands on the same search re-run against the live feed.

No. The feed has to be reachable over public http or https. Feeds behind authentication, personalised premium feeds tied to a subscriber token, and feeds on private or internal network addresses are all rejected by the server-side fetcher, which blocks non-public hosts. For a private feed, use whatever search your podcast host or listening app provides.

Those platforms search their own indexed copy of a show's metadata, built on their own refresh schedule and sometimes normalised or truncated along the way. This tool reads the live feed at the moment you search, so it reflects what the publisher is serving right now — including an episode published minutes ago that a directory has not ingested yet.

Finding the episode where a specific guest, book, product, or place was covered; checking whether you already made an episode on a topic before pitching or recording it again; auditing your own show notes for terms you assumed you had covered; and pulling a dated list of episodes mentioning something for a newsletter or a link round-up. Because it reads the publisher's own feed, the titles, dates, and links it returns are checkable against the source.