← All guides

How to transcribe a podcast episode

How to transcribe a podcast episode

Most podcasters know they should have transcripts. Fewer actually do it, because it sounds like a chore on top of an already long production process. It doesn't have to be. With a decent tool, a transcript is a ten-minute job, and once you have it, it quietly does a lot of work for you.

Here's how to do it, and more importantly, what to do with it afterwards.

Why bother transcribing a podcast?

Let me skip the generic benefits and give you the ones that actually show up in practice.

Search engines can't listen. Google reads text. An episode with a transcript gives it hundreds or thousands of words to understand, including all the specific phrases your guest used. That's how an episode about a niche topic ends up found by someone searching for exactly that topic.

Not everyone can or wants to listen. Some people are deaf or hard of hearing. Some are at work with no headphones. Some read faster than you talk. A transcript lets all of them in.

You get content for free. Show notes, blog posts, newsletter sections, social quotes and video captions all start from the same place: the words you already said.

It makes you better. Reading your own transcript is humbling and useful. You'll notice the filler words, the rambling intros and the questions that could have been shorter.

Before you transcribe: export the right file

You want the cleanest audio you have, so use your final export, not a copy that's been passed around.

  • If you have the multi-track recording, you can transcribe the final mixed episode. Don't bother with individual tracks unless you need separate speakers.
  • Skip the intro music and ad reads if you can. Trimming a copy of the file that starts at the first word of the conversation gives you a cleaner transcript, since music often produces strange phantom words.
  • MP3, M4A and WAV all work. If you've got a choice, WAV or a higher-bitrate file is better than a heavily compressed one. Most podcast files are MP3, which is fine. See MP3 to text if that's what you have.

If your episode is on video too, such as a YouTube version, that works as well. The tool only reads the audio track.

Step 1: Run the transcription

Open the audio to text tool, drop in the episode and start it. A long episode takes a while, particularly in the free mode, since it runs on your own device. Keep the tab open and let it work. A laptop or desktop is much more comfortable than a phone for an hour-long episode.

If you're not sure which model size to use, our guide to Whisper tiny vs base vs small explains the trade-offs. For a podcast with clear voices, a mid-size model is usually the best balance between speed and accuracy.

Step 2: Fix the things software always gets wrong

You don't need to proofread every word to a publishing standard, but a quick pass pays off. Podcast transcripts have a few predictable weak spots:

  • Guest and company names. Check the spelling against their website or social profile. Misspelling a guest's name in public is the one mistake they'll actually notice.
  • Product names and jargon. Niche terms get turned into similar-sounding ordinary words.
  • Numbers and statistics. If you cited a figure, make sure it's right in the text.
  • Sponsor reads. Brand names and discount codes need to be exact. Correct them or remove the section.
  • Laughter and crosstalk. Software often turns laughter into odd words. Delete them or replace with "[laughs]".

Do this with the audio playing at a slightly faster speed. You'll catch most errors without feeling like you're listening to the whole show again.

Step 3: Add speaker names

Most tools give you plain text without telling you who is speaking. For a two-person show, you can add labels in a few minutes: start a new line at each change of speaker with their first name.

Priya: Welcome back. Today we're talking about pricing.

Dan: Which everyone hates, so it's perfect.

For a solo show, you can skip this entirely. For a panel with four or five voices, decide whether the effort is worth it. A transcript with no speaker labels is still useful for search and quoting, and it's honest to say so at the top of the page.

Step 4: Turn it into something readable

A raw transcript is a wall of text. Before you publish it, give readers something to hold onto.

  • Break it into short paragraphs. One idea per paragraph.
  • Add subheadings every few minutes, based on the topic changes. Skim the audio and write a short, specific heading for each section.
  • Add timestamps at topic changes so readers can jump to that point in the audio.
  • Add a short intro paragraph summarizing who the guest is and what the episode covers.

You don't need to make it fancy. A clean, scannable page is already better than most podcast transcripts out there.

What to do with the transcript

Now the interesting part. One transcript can become several pieces of content.

Show notes

Skim the transcript and pull out the main topics, the resources the guest mentioned, and a few key takeaways. Show notes written from a transcript are more accurate than ones written from memory, and they take a fraction of the time.

A blog post

Pick the best ten minutes of the conversation and rewrite them into an article. Don't just paste the transcript and call it a post. Reorganize it around a clear question or idea, tighten the wording, and keep the guest's best lines as quotes. You'll get something that stands alone for readers who never listen to the episode.

Quote graphics and social posts

Search the transcript for the line that made you laugh or think. Copy the exact wording, double check it against the audio, and you've got a quote for social media. Having the text means you can find it in seconds, instead of scrubbing through an hour of audio.

Captions for video clips

If you post short clips on social media or YouTube, you need captions, because many people watch with the sound off. Export the transcript with timing information as an SRT or VTT file and use it as captions. We cover that in our guide to turning audio into SRT subtitles.

Newsletter content

Three takeaways from the episode, plus one great quote and a link, makes a very easy newsletter section. It's much easier to write when the exact words are in front of you.

Where to publish the transcript

You've got a few options, and each works:

  1. On the episode page, below the show notes. The simplest option. Everything is in one place, and the text helps that page get found.
  2. On a separate transcript page, linked from the episode. Good if the transcript is very long and you don't want to clutter the main page.
  3. As a downloadable file. Fine for accessibility, but less helpful for search than text on the page.

If you publish on the page, use real text rather than an image or a PDF only. Text is what people and search engines can actually read.

One note: don't copy the same transcript text onto several pages of your site. Pick one home for it.

A word on accuracy and honesty

An automatic transcript isn't perfect, and you don't have to pretend it is. If you haven't proofread every line, a simple note such as "This transcript was generated automatically and lightly edited" is honest and helps readers forgive small mistakes. If accuracy matters a lot, for example for a legal or medical podcast, do a full pass.

How long does it take?

For a typical 45-minute episode with clear audio, a reasonable workflow looks like this:

  • Transcription: a few minutes to maybe twenty, depending on your device and the model.
  • Cleanup of names and obvious errors: 15 to 20 minutes.
  • Speaker labels and headings: 10 to 15 minutes.

That's under an hour for a transcript you can publish, plus raw material for show notes, a post and a handful of social quotes. Compare that with typing it yourself or paying per minute for every episode, and the free route makes a lot of sense.

The short version

Export your cleanest audio, run it through a speech-to-text tool, fix names and numbers, add speaker labels and headings, then publish it and reuse it everywhere. The transcript isn't just an accessibility add-on. It's the raw material for most of your content.

Want to try it on your latest episode? Open the audio to text tool, drop in the file and see how much of the work is already done.