supermemoryai
[Feature Enhancement] Use Mozilla Readability for content extraction
Firefox reader mode simplifies many websites by extracting just the content, the library that does this is open source: https://github.com/mozilla/readability
Page extraction could use this to better target page content. If the backend can accept content directly (Instead of trying to go extract it itself), it's conceivable that the browser extension could use this as a method to extract that content early.
@Dhravya Are you accepting pull requests and enhancements? If so, what scope is acceptable for those changes? Some enhancements may not be integrate-able without structural changes. 🤔
We accept any kind of pull requests. Structural changes can be discussed here!
This is very helpful. One problem we had was reliable extraction of text from the website and tbh, we just gave up on it.
This would actually be very helpful.
Can this also be used by our browser rendering agent? Shouldn't be too big of a change, readability looks like a one line API
We can also use this https://r.jina.ai/dhravya.dev
Text extraction is difficult, this is one of the few tools I've found that do it somewhat reliably. I was doing light work on a personal knowledge archiver (Automatically doing text extract of webpages I visit with archivebox archiving for later vecorization and second-brain retrieval). You can probably tell why your projects excites me so much :)
haha! Thanks douglas!
I am looking at JinaAI and FireCrawl by Mendable https://x.com/mendableai/status/1780289422644109686 It looks like most are just using turndown, Firecrawl is even server side so it doesn't work with sites that have captcha etc.
Readability and Jina are genuinely the only two good solutions i could find
The task is to basically call the function in https://github.com/mozilla/readability while getting content.
Both in the extension and the getPageContent API route.
That's awesome!
I'm happy you took the suggestion thank you. I wish I had the time to contribute right now.