How to Get a Transcript of Any Podcast (Every Method That Works in 2026)

Somewhere around minute 40 of a two hour episode, someone says the thing. The line you want to quote in a doc, send to a friend, or check against reality later. And then the episode keeps going, and the line is gone, buried in an audio file with no Cmd+F.

This is the moment most people go looking for a transcript. If that’s you, good news: in 2026 there is a free way to get the text of almost any episode in under two minutes. The bad news is that the methods are scattered across four different apps and nobody tells you which one to try first.

I’ve tested all of them. Here’s the full ladder, from “it already exists, just look” to “generate it yourself with AI,” plus an honest note at the end about when a transcript is the wrong tool entirely.

Step 1: Check whether the transcript already exists (it usually does)

For most shows of any size, someone has already made the transcript. Four places to look, in order of how often they pay off.

Apple Podcasts

Since iOS 17.4, Apple Podcasts generates transcripts for nearly every episode in its catalog, in multiple languages. Open the episode, tap the speech bubble icon in the player, and the full transcript appears, highlighting word by word as the audio plays. You can search inside it and tap any line to jump the audio to that moment.

The catch: the transcript lives inside the app. There’s no export button, and getting text out is clumsy. It’s excellent for reading along and finding a moment, not for quoting at length.

Spotify

Spotify auto-generates transcripts for a large and growing share of its catalog. On the episode page, scroll down past the description and you’ll find the transcript section where it’s available. Same deal as Apple: synced, searchable, read-only. Good for following along, no way to take the text with you.

The show’s own website

The bigger the show, the more likely it publishes real transcripts, with speaker names and human-grade accuracy. Acquired posts complete transcripts for every episode on its site. Dwarkesh publishes his interviews as full transcripts on his Substack. Tim Ferriss has posted transcripts for years. Lenny’s Podcast episodes ship with transcripts through the newsletter.

The move: search the show name plus the episode title plus the word “transcript.” When this works, it’s the best version you can get. Copyable, formatted, and someone actually checked the names.

YouTube

Most podcasts now upload full episodes to YouTube, and every YouTube video has a transcript panel. On desktop: expand the video description, click “Show transcript,” and a timestamped panel opens on the right. The three-dot menu at the top of the panel toggles timestamps off, which gives you clean text you can select and copy.

This is the fastest free way to get copyable text for a show that doesn’t publish its own transcripts. Accuracy is auto-caption grade: fine for finding and quoting, occasionally creative with names and technical jargon.

Step 2: If it doesn’t exist, make one with Whisper (free)

When an episode is too small, too new, or too niche to have a transcript anywhere, you can generate one yourself, and the tool for that is settled at this point: OpenAI’s Whisper. It’s open source, it’s free, it runs entirely on your own machine, and on clear speech it’s close to human accuracy.

The easiest way to use it on a Mac is MacWhisper. You drag an audio file in and get a full transcript with timestamps, exportable as plain text, markdown, or subtitles. The free tier’s smaller models are plenty for podcast audio; the paid tier unlocks the larger, more accurate ones.

The only fiddly part is getting the audio file, because podcast apps hide the actual mp3. Every show has a public RSS feed, and the feed contains a direct link to each episode’s audio. Find the feed through the show’s website or any podcast feed lookup tool, open it, and copy the audio URL for your episode. Or skip all of this: if the episode is on YouTube, the transcript panel above already solved your problem.

If you live in the terminal, whisper.cpp transcribes an hour-long episode in a few minutes on Apple silicon with a single command. Same model, no app.

Step 3: Paid tools, if transcripts are a weekly habit

If you’re transcribing regularly, a few tools earn their subscription:

• Snipd is a podcast player with transcripts built in for everything you listen to, plus AI chapters and highlight capture. If what you want is “transcripts included” as a listener rather than a producer, this is honestly the best app-native answer. (I wrote a full Margin vs Snipd comparison if you’re weighing them.)
• Descript is transcription plus an audio editor where you edit the audio by editing the text. It’s built for people who make podcasts, not just listen to them.
• Rev sells human transcription at a few dollars per minute. You use it when accuracy has legal or journalistic stakes and “close to human” isn’t enough.

The whole ladder in one table

The honest question: do you actually want a transcript?

Here’s the thing I noticed after years of doing this. Watch what you actually do with a transcript once you have it: you Cmd+F your way to the one moment you half-remember, read 40 seconds’ worth of text, and close the tab. You didn’t want 12,000 words. You wanted the 40 seconds.

A transcript is retrieval after the fact. It’s the expensive way to find a moment, because you’re reconstructing at midnight what you already knew at minute 40: this part matters.

The cheap way is to mark the moment when it happens. That’s the entire idea behind Margin: you press and hold while listening, say what you’re thinking, and the note transcribes on-device and pins itself to the exact second of the episode that triggered it. Later, you don’t search a wall of text. You scroll your own moments.

Both tools have real jobs. When you need the full text (research, accessibility, quoting at length), use the ladder above; it’s never been easier. When what you really need is your moments back, capture them as they happen, and you’ll stop needing transcripts for the thing transcripts are worst at.

For the capture side of that equation, I wrote a companion piece: how to take better notes while listening to podcasts.

Selinay