YouTube Transcript Cleaner and Reader Tool

Hacker News

Share card

YouTube Transcript Cleaner and Reader Tool

I made this little self-hosted tool to scratch an itch I had-- I often want to check out long YouTube videos, but they are very slow and I can read much faster than I can listen to something. But I couldn't find a good way to do this without a lot of effort and waiting. I know there are various tools that use OpenAI's Whisper to generate transcripts for YouTube videos, but they take forever to run and the output isn't very nice to read. YouTube already has quite good automatic transcripts for most videos, but the problem is that they are also not really formatted to read easily. I discovered that there is a good free Chrome extension to extract the official YouTube transcript as text to your clipboard (link to the extension is in the readme in my repo), but that just gets you plain text with awkward line breaks. So this little html file can be saved to your local computer and opened in Chrome. Then you just go to the video you want, get the entire transcript in your clipboard using that Chrome extension, and paste it into my tool and click the button to clean it. It takes a couple seconds to run (since it's using a heavy library to determine the optimal line breaks) but then you can view it in a nice little reader mode that makes it quite pleasant. Hopefully people will find it useful. Oh, and the most interesting part of all this from my perspective is that ChatGPT wrote basically all of it from scratch in around an hour or so-- I only had to glue together some code and give feedback. I shared a link to the ChatGPT conversation in the readme file so you can see how I was able to do this step by step.