Creator-provided podcast transcripts with speakers in Metacast mobile app

TL;DR

Earlier this year, we started to show transcripts provided by creators in the web app. We're now bringing creator-provided transcripts to the mobile app on iOS and Android.

• When podcasters add a transcript to their RSS feed, Metacast will use it instead of transcribing audio with AI.
• You can now see speakers in transcripts when speakers are provided by the creator.

What's changing?

We added the under-the-hood functionality to use transcripts provided by podcasters in the podcast feed. Some of those transcripts have speakers, and when they do, magic happens.

Here's an example of The Panel podcast, hosted by our friend Justin Jackson from Transistor.fm and Brian Casel. The RSS feed for The Panel includes the transcript with speakers, so we now use it in Metacast to show who's speaking.

On the web, we're also showing speaker pictures when they're provided in the feed. This feature is coming to the mobile at some point too.

What is a "creator transcript"?

When podcasters publish their podcast on a platform that supports transcripts, like Transistor, they have the option to supply a transcript to go along with the episode.

The platform includes a podcast:transcript tag (docs) to the RSS feed. Podcast apps subscribing to the feed can read the transcript and display it to listeners.

How to transcribe a podcast episode

• Manual transcription by humans — arguably, the highest quality transcript ever. Also, it is the most expensive method.
• Transcription by the podcast hosting platform — when an audio file is uploaded, it gets automatically transcribed by the hosting platform. The episode from The Panel podcast above was transcribed with Transistor's AI Transcription feature.
• Transcription by podcast editing software — if you edit your podcast in Descript or Riverside, you can export the transcript and upload it with your episode.
• Transcription by a specialized podcast transcription service — shameless plug... we have a service "Metacast for Podcasters" in private beta. Reach out to us at support@metacast.app if you want to learn more and get early access.
• Transcription by a general-purpose speech-to-text service — you can use tools like SpeechIsCheap from our friend Ilya K. (we use it for Metacast!), OpenAI's Speech-to-Text API, or a similar service to transcribe audio. This process is a bit more involved, because you'll need to convert the transcript from the Whisper format to the format expected by podcast apps.

Podcast transcript formats

If you want podcast apps to use your transcript, you'll need to provide it in one of the specialized transcript formats (or ideally in all three):

• JSON — the most structured transcript format (spec).
• WebVTT — the Web Video Text Tracks format was originally created to caption video content. Now it's also used in audio podcasting.
• SRT — the SubRip Text format is also a video captioning format and it's similar to WebVTT.

You can also provide transcripts in non-standard formats, like HTML or plaintext. However, most podcast apps will ignore them, because they don't know how to parse them.

How do I try this?

The transcript feature is available to Metacast Premium users. Starting today, all Premium users will start seeing creator-provided transcripts when they're available (for episodes where we've not yet generated a transcript with AI).

If you're seeing a transcript with speakers, you can be sure that it's a creator-provided transcript.

Note for podcasters

If you want to verify whether your transcript is displayed properly, check out this FAQ.

P.S. Also available as a podcast

You can also listen to us talk about creator-provided transcript on the Metacast: Behind the Scenes podcast episode 71. Survaviversary: We survived 1 year in production + limited time sale for new users.