How to find a podcast episode by describing it

On a run a few months ago you heard a guest explain why sourdough starters behave differently in winter, or how a small company priced its first product, or a story about a lighthouse keeper that you’ve wanted to retell since. You subscribe to fifteen shows. The episode could be any of the last few hundred, and podcast episodes run long, so even with the right one in hand, the story is somewhere in ninety minutes of audio.

This page covers how people usually try to find a podcast episode they half remember, and how to save episodes to Fabric so they’re transcribed and searchable by what was said, with results that open at the right moment.

How people usually look for a podcast episode

Podcast apps are the first stop. Apple Podcasts, Spotify and Overcast keep a listening history and let you search episode titles and descriptions. That works when the topic made it into the title or the show notes. Guests often tell their best stories in passing, so the moment you remember may not be mentioned anywhere outside the audio itself.

A web search is the next try, typing something like “podcast guest lighthouse keeper story”. For popular shows, someone may have written about the episode or posted a clip. For smaller shows, results are thin. Some shows publish transcripts on their websites, which makes this easier where they exist.

Some listeners keep a note on their phone with episode names and rough timestamps (“Ep 212, 48 min, lighthouse”), or take a screenshot of the player at a good moment. Both work well when the habit holds. They capture a pointer to the moment, and you still need to remember to write one down before the episode moves on.

Once there are a few candidate episodes, the remaining work is listening. People skip ahead in chunks, listen for a familiar voice, and skip again. For a two-hour interview, that can take most of an evening, and it’s easy to skip straight past the part you wanted.

Step 1: Save episodes as you listen

In most podcast apps, the episode screen has a Share button. Tap it on iOS or Android and choose Fabric to save the episode link. If you don’t have the app yet, it’s on the mobile download page. On a computer, copy the episode’s link and paste it into Fabric, where it becomes a bookmark.

Saving takes a few seconds, and the habit pays off: an episode that made you think, even briefly, is worth a save. You can add a comment at the same time, such as “pricing story near the end”, which gives search another clue later.

If you have audio files instead of links, such as downloaded episodes or a show you produce yourself, drag the MP3, WAV or M4A files into Fabric or click Upload a file.

Step 2: Let Fabric transcribe the episode

Fabric transcribes podcast links directly, along with uploaded audio and YouTube and TikTok links. The transcript becomes searchable, you can edit it if a guest’s name comes out wrong, and the AI assistant can answer questions about it with timestamps. Transcription needs a paid plan.

Fabric also creates embeddings for each episode so search matches by meaning, writes a summary, extracts topics and suggests tags. With smart organization you can set up smart collections, such as one for business interviews or one for food and cooking, that fill themselves as episodes arrive.

Step 3: Search for what the guest said

Click Search or press Cmd+F (Mac) / Ctrl+F (Windows) and describe the moment. Fabric search matches by meaning across transcripts, so “why sourdough starter is sluggish in cold weather” can find an episode where the baker said “in January the culture slows right down because the kitchen’s ten degrees cooler”.

Open the result and the episode opens at the timestamp where that passage sits. You can listen to the story itself, then carry on or go back a few minutes for context.

Specific, concrete details make the best queries: an object, a place, a number, an unusual comparison. If you remember a name, a company or a book title, put it in quotes for a strict match. Filtering by kind of item keeps results to saved links and recordings, and tags narrow it to one show if you’ve tagged them.

Step 4: Ask the AI assistant across episodes

Some questions span several episodes: “Which guests have talked about pricing a first product?” Ask the AI assistant and it answers with clickable references to each episode, including timestamps for each point it draws on. Follow-up questions help narrow it down (“Only the ones from this year”).

Keep a show’s episodes in their own folder and open the assistant inside it to limit it to that show. Answers can be saved as notes, which is a useful way to build a page of ideas from a series over time. When you find one episode you liked, select it and run similarity search to bring up episodes like it from your saves.

Frequently asked questions

How do I find a podcast episode when I only remember what was said?

Save episodes to Fabric so they’re transcribed, then search for what you remember. Results open at the matching timestamp.

Can Fabric transcribe podcasts from a link?

Yes. Fabric transcribes podcast links, as well as uploaded MP3, WAV and M4A files. Transcription needs a paid plan.

How do I save a podcast episode from my phone?

Tap Share on the episode in your podcast app and choose Fabric.

Can I jump to the moment something was said?

Yes. Search results and AI answers link to the timestamp in the episode.

Can I fix a guest’s name in the transcript?

Yes. Transcripts can be edited, which helps when a name or technical term was misheard and you want to search for it later.

Can AI summarize what several podcast guests said about one topic?

Yes. Ask the AI assistant, limited to a folder of episodes if you like, and it answers with timestamped references you can click.

Is podcast transcription free?

No. Transcription of podcast links and audio files needs a paid plan. See pricing for details.

Can I find episodes similar to one I liked?

Yes. Select the episode and run similarity search to find related items among your saves.

Can I search podcasts and my other notes at the same time?

Yes. One search covers transcripts, notes, documents and saved links together, so an episode can turn up next to an article on the same subject. How to find something you saved months ago has more techniques for older saves.