New v0.6 — Record meetings locally · Korean, Japanese & 96 more languages. See what's new →

Podcasts → Text · Local · Free

Transcribe a podcast episode to text — on your Mac

Paste an Apple Podcasts link — or just search by name — and get a clean transcript on your Mac. Local, private, and no per-minute fees, so a whole back catalog costs nothing but a little compute time.

Paste a link like https://podcasts.apple.com/…?i=…

macOS 14.2+ · Apple Silicon · Free & open-source · no account

Why transcribe a podcast?

A podcast is hours of dense, quotable conversation locked inside audio. A transcript unlocks it: you can skim an hour in ten minutes, quote a guest word-for-word, and search a back catalog you could never re-listen to.

  • Researchers & students — turn an interview into a citable source with exact quotes and timestamps.
  • Journalists — pull the precise wording of a statement instead of paraphrasing from memory.
  • Podcasters — generate show notes, chapter markers, and social clips from the text in minutes.

MacParakeet resolves an Apple Podcasts link to its audio through the public iTunes lookup — or you can skip the link hunt entirely and search by name from the CLI. Either way it runs on your Mac, with no per-minute fees, so transcribing a 200-episode archive is free. For a step-by-step walkthrough with batch commands, see how to transcribe a podcast on your Mac.

How to transcribe podcast episodes on your Mac

Free, local, and no upload to a transcription cloud. The whole thing takes about a minute the first time, seconds after that.

  1. 1

    Get the episode link — or just its name

    Copy an Apple Podcasts episode link (podcasts.apple.com/…?i=…), a direct MP3 from the show's site, or note the show and episode title. Any public RSS feed works.

  2. 2

    Paste it — or search from the CLI

    Paste the link into the Transcribe tab, or from the terminal run macparakeet-cli transcribe --podcast "show + episode" to find it through the iTunes directory and its RSS feed. MacParakeet resolves it to the audio and transcribes locally.

  3. 3

    Notes, quotes, and a searchable archive

    Export Markdown for notes, pull exact quotes with timestamps, or transcribe an entire season into a folder of searchable text you keep forever.

Prefer the terminal? One command does the download and the transcription:

$ macparakeet-cli transcribe --podcast "Lex Fridman episode 400" --format text

Install with brew install moona3k/tap/macparakeet-cli. Add more links and --output-dir to batch a whole list in one walk-away run. See the agent & CLI guide →

Already downloaded the clip, or just need a minute or two? You can also transcribe a short file right in your browser — no install, nothing uploaded.

What "local" means here — honestly

To fetch a podcast link, MacParakeet downloads the audio from the source over the network — that step needs an internet connection. Everything after that is on-device: the speech recognition runs on your Mac's Apple Neural Engine via the Parakeet model, at roughly 155× realtime with a ~2.5% word error rate on English. Your audio is never uploaded to a transcription server, there's no account, and there are no per-minute caps.

Transcribe content you own or have the right to use, and respect each platform's terms of service. MacParakeet is a tool for getting text out of media you're allowed to access — not for redistributing other people's videos.

Frequently asked

Can I transcribe Spotify podcasts?

Public episodes with an RSS or Apple Podcasts link work. Spotify-exclusive or DRM-protected shows have no public download path, so they can't be transcribed this way — stick to shows with a public RSS or Apple Podcasts feed.

Can I search for an episode instead of hunting for a link?

Yes. From the CLI, run macparakeet-cli transcribe --podcast "Lex Fridman episode 400". It searches the iTunes directory, reads the show's RSS feed, and picks the episode by your hints (or the latest).

Is there a per-minute cost like cloud transcription?

No. It's free and runs locally, so a 200-episode back catalog — which can cost hundreds of dollars on metered cloud services — costs nothing but a little compute time.

How accurate is it on a two-mic podcast?

Very accurate on clean, two-mic recordings — around 2.5% word error rate on English. Heavy music beds, phone-quality guests, or low-bitrate audio lower accuracy. Built-in speaker diarization labels who said what.

Get the transcript in a minute.

Free and open-source. Download MacParakeet, paste your podcast link, and keep the text.

macOS 14.2+ · Apple Silicon · GPL-3.0