Podcast transcription sits in a weird spot in 2026. Half the tools that come up when you search it are free (and better than they have any right to be). The other half are paid products charging professional-services rates for output that is now essentially a commodity. And a third category doesn't produce transcripts at all — they produce summaries or briefings built on top of transcription, which is usually closer to what people actually wanted in the first place.

This guide walks through the real categories, the tools worth using in each, and the quiet question most transcription searches don't answer: do you actually want a transcript, or do you want something that happens to be built out of one?

The four real categories of podcast transcription

Category 1 — Platform auto-transcripts (free, already done for you)

The biggest change in the last two years: Apple Podcasts, Spotify, and YouTube all auto-generate transcripts for a large chunk of the shows in their catalogs. If the show you care about is distributed through any of those three platforms, a transcript probably already exists and you don't need to pay for one.

Apple Podcasts — Auto-transcripts for most shows in the catalog as of 2024. Tap the speech-bubble icon inside the episode player on iOS or macOS. Time-synced, searchable, but hard to export cleanly.

Spotify — Coverage is growing but patchy. Big shows have transcripts, many independent shows don't. The transcript shows up under the player controls on mobile when it's available.

YouTube — Auto-captions for any uploaded video, and a surprising number of podcasts now also publish on YouTube. Click the three-dot menu under the video and pick "Show transcript." Copyable text panel, free, no watermark.

Cost: Free. Quality: Usually decent for English with clear audio. Catch: Poor speaker attribution, hard to export in bulk.

Category 2 — Open-source transcription (free, best quality, more effort)

OpenAI Whisper is the reason professional transcription services had to reinvent themselves. It runs on your computer, has no API quota, no watermark, no file-size limit, and produces output that rivals human-edited transcripts for English with clear audio. The catch is that you need to be comfortable with the command line and willing to wait while it processes.

How to use it: Install via pip install openai-whisper, download the MP3 of the episode, run whisper episode.mp3 --model medium. A 60-minute episode takes 10 to 30 minutes to process depending on your hardware. Outputs plain text, SRT subtitles, VTT, or timestamped JSON.

Cost: Free. Quality: Very good for English; speaker diarization requires a separate tool. Best for: Researchers, journalists, and people who transcribe more than a few episodes a month.

Category 3 — SaaS transcription tools (paid, with free tiers)

Otter, Descript, Rev, Trint, Podscribe, and similar services. All produce reliable transcripts with speaker attribution, searchable interfaces, and export options. Free tiers cap you at 30 minutes to a few hours per month — useful for one-offs, not for ongoing work.

When they're worth paying for: You need speaker-labeled transcripts regularly, you want integrated editing (especially Descript), or you need team access to a shared transcript library.

When they're not: You transcribe one or two episodes a month. At that volume, Apple Podcasts transcripts or Whisper cover it free.

Category 4 — Transcription as plumbing (you want the summary, not the text)

This is the category most transcription searches belong in and don't realize it. If the reason you want a transcript is "so I can remember what the episode said" or "so I can find the interesting parts without listening to the whole thing," a transcript is the wrong deliverable — it's an 8,000-word wall of text you will skim once and never read again.

What you actually want is the summary that would have come out of the transcript. Text summary tools (Podsqueeze, Podnotes) build on transcription under the hood and give you back bullets. Audio-briefing tools (TrimCast, which is us) build on transcription under the hood and give you back a listenable short version of the episode in the same audio format you wanted in the first place.

How to pick the right podcast transcription path

Match the method to what you'll actually do with the output:

  • One-time transcript of one episode → Apple Podcasts or YouTube auto-captions, free, zero setup
  • Research across many episodes → Whisper locally, or Otter/Descript paid tier
  • Legal, journalistic, or quotable accuracy required → Rev (paid human transcription) or Descript with manual editing
  • "I don't actually want to read 8,000 words, I just want to know what the episode was about" → text summary tool (free tier of any of them)
  • "I want to listen to the episode's highlights in the time I'd spend reading a summary" → audio briefing tool like TrimCast. Paste the URL, pick Quick Brief (10 min), Essential (15–20 min), or Deep Cut (25–45 min), and you get a multi-voice audio recap that covers the same ground as the full episode.

The honest answer

For 90% of people who search "podcast transcription," the combination of Apple Podcasts auto-transcripts (for reading) and an audio briefing tool (for listening) covers every real use case for free or close to free. The paid transcription SaaS category makes sense if you have a specific professional workflow — research, journalism, legal, content production — but for casual listeners it's a solution to a problem the free tools already solved.

Start there. Upgrade only if the free options miss what you actually need.

Your AI podcast assistant. Listen smarter, not longer.