Podcast Text Downloader Idea

#1

Post by mpichardo » Wed, Feb 22, 2023, 7:15 PM UTC

It would be awesome if you went to a website adn copy and pasted a link to a podcast and then you can get an output of a text transcript of the podcast.

I found a hacky way of doing this by using a youtube mp3 downloader and then then uploading that file into otter.

but was considering building an app that automated this process. The next step would be to leverage otter or gpt-3 to give you a condensed summary of this podcast.

Is this something that is legal? Is this something that people would pay for?

Cheers, MP

Re: Podcast Text Downloader Idea

#2

Post by blueberrychpstx » Wed, Feb 22, 2023, 8:32 PM UTC

I’m doing this as part of the Solana hackathon with some extra bits.

We’ll share here if I can remember!

Use whisper for transcription. I have paid for Otter for years and am a little work away from replacing it with my own open source web socket based solution to phone home to whisper running on desktop

Also check out Deepgram for programmatic transcriptions with diarization.

Re: Podcast Text Downloader Idea

#3

Post by jamescridland » Thu, Feb 23, 2023, 2:39 AM UTC

Arguably, this is a transformative work and would require the permission of the creator.

steno.fm is an app that surfaces transcripts from podcasts who are doing their own transcriptions and placing them in the RSS feed.