Transcribe Podcast to Text: Easy No-RSS Guide 2026

Short answer: To transcribe podcast to text, you need two things: the episode’s audio and a speech-to-text model. The model is the easy part — OpenAI Whisper is accurate and costs about $0.36 per hour of audio. The hard part is the audio. Most tools want you to upload a file or paste an RSS feed URL, and the podcast app you actually listen in gives you neither. The fastest route is a tool that lets you search for the episode by name and hands back a .txt or .srt file.

I ran into that gap often enough that I built Podskrift around it. This guide shows the three ways to transcribe podcast to text, when each one makes sense, and what I learned about getting complete, accurate transcripts along the way. Disclosure: Podskrift is my own product.

Why most ways to transcribe podcast to text stall on RSS and uploads

Podcasts are distributed as RSS feeds. Every show you hear in Spotify or Apple Podcasts is, underneath, a feed with an audio file attached to each episode — it’s literally in Apple’s podcast requirements. Transcription tools are built around that fact, so they ask for one of two inputs:

• An audio file to upload. But podcast apps don’t give you files. Downloads in Spotify and Apple Podcasts are for offline listening inside the app.
• An RSS feed URL. The feed exists, but listening apps don’t show it. You end up digging through the show’s website or a feed directory.

The third option people try is the built-in transcript in their podcast app. Those are reading views without an export button — I covered that in detail in the guide to getting a Spotify podcast transcript you can export.

So when you want to transcribe podcast to text, the friction usually isn’t transcription at all. It’s getting from “I heard this episode” to the audio.

Three routes, compared

There are three ways to transcribe podcast to text, and they differ mainly in what you have to bring:

Route	You need	Best for	Watch out for
Upload an audio file	The MP3 on your computer	Your own recordings	Podcast apps don’t export files
Paste an RSS feed	The show’s feed URL	Shows outside search catalogues	Finding the feed takes time
Search by name	The show or episode title	Anything you heard in an app	Spotify-exclusive shows have no public feed

If you host the podcast yourself, uploading works fine — you already have the file. For everything else, searching by name is the quickest way to transcribe podcast to text, and the RSS feed is the fallback when a show isn’t in the catalogue. Podskrift has a short guide on finding a podcast’s RSS feed for those cases.

How to transcribe podcast to text in 5 steps

Here’s the whole process to transcribe podcast to text with Podskrift. There’s no file to upload and no feed to hunt for.

1. Search. Open podskrift.com. Search by Podcasts to browse a show, or by Episodes to jump to a specific one.
2. Pick the episode. Each episode shows its length and an estimated cost before you start.
3. Set the language. Leave it on auto-detect, or name the spoken language — more on when that matters below.
4. Transcribe. You can watch progress live, and the job keeps running if you close the page.
5. Download. Grab the .txt file, the .srt file, or both.

New accounts get 60 minutes of audio free, so your first episode or two costs nothing. A typical one-hour episode takes a few minutes to process.

.txt or .srt: what you actually need

When you transcribe podcast to text, both formats contain the same words. The difference is timing.

• .txt is plain continuous text. Use it for reading, searching, notes, and pasting into ChatGPT or Claude to summarize an episode or ask questions about it.
• .srt (SubRip) splits the text into numbered segments with start and end timestamps. Use it when you need to find the exact moment something was said, cut a clip, or add subtitles to a video.

If you’re unsure, download both. The .srt is also the quickest way to check a transcript is complete: the last timestamp should match the length of the episode.

Pick the spoken language (and when not to)

Podskrift supports 28 languages, from English, Spanish and Mandarin to Norwegian, Ukrainian and Vietnamese. The underlying Whisper model was trained on many more.

The language setting taught me something the hard way. An early version of Podskrift defaulted to Norwegian. A listener transcribed a Japanese episode without changing it and got phonetic nonsense back. Whisper doesn’t treat the language you give it as a hint — it treats it as a constraint, and forces the audio into that language.

So the rules are simple:

• Leave it on auto-detect for most episodes. Whisper identifies the language from the audio.
• Name the language for short clips, heavy accents or shows that mix languages. That’s where auto-detect is most likely to guess wrong.
• Never pick a language the episode isn’t in. You’ll pay for a transcript that’s unreadable.

This is especially useful for smaller languages. If you want to transcribe podcast to text in Norwegian, Danish or Finnish, naming the language removes the risk of auto-detect guessing a neighbouring language from the first few sentences.

What happens under the hood

You don’t need to know this to transcribe podcast to text, but it explains why some transcripts come back incomplete — with any tool.

OpenAI’s speech-to-text API accepts files up to 25 MB, and a long episode is far bigger than that. So Podskrift converts the audio to 16 kHz mono and splits it into 15-minute pieces before sending anything to Whisper.

Two things went wrong before I got that right, and both are worth knowing:

1. Splitting by file size doesn’t work. Podcast audio is often variable bitrate, so bytes aren’t proportional to time. One episode with a dense first half produced a 27 MB piece against a 24 MB target — over the API limit.
2. Reported length can lie. Many podcasts insert ads dynamically, which stitches the file together from segments. On one of those files, the reported duration was shorter than the real audio, and an early version skipped 16 minutes without raising a single error.

The fix for both was to walk the actual audio stream when splitting, instead of trusting the file’s metadata. Whichever tool you use to transcribe podcast to text, check the last timestamp in the .srt against the episode length.

What it costs to transcribe podcast to text

Podskrift doesn’t have a subscription. You get 60 free minutes on signup, then add your own OpenAI API key and pay OpenAI directly at $0.006 per minute:

Episode length	Approximate cost
30 minutes	$0.18
1 hour	$0.36
2 hours	$0.72
3 hours	$1.08

That’s the whole bill to transcribe podcast to text — no seats, no minimum. For a longer comparison with per-minute transcription services, see my full cost breakdown. The Podskrift product page has the rest of the details.

Transcribe Podcast to Text FAQ