Text-Only NPR

jumploops wrote: post #3

Perfect for LLMs! We use a html->markdown converter for web scraping and sites like this make it even easier/more robust. Side note: bring back the RSS feeds?

Curious, can you recommend a HTML -> MD converter?

Re: Text-Only NPR

#5

Post by sevensor » Thu, Nov 07, 2024, 1:27 PM UTC

alborzb wrote: post #2

This reminds me of the CNN Lite website -- https://lite.cnn.com/

Both of which I prefer to the ordinary versions. So much less noise, much slower pace. Only 80% junk, which is super low by the standards of modern news websites.

Re: Text-Only NPR

#6

Post by zem » Thu, Nov 07, 2024, 7:01 PM UTC

pinjasaur wrote: post #4

jumploops wrote: post #3

Perfect for LLMs! We use a html->markdown converter for web scraping and sites like this make it even easier/more robust. Side note: bring back the RSS feeds?

Curious, can you recommend a HTML -> MD converter?

pandoc is pretty good https://pandoc.org/demos.html

Re: Text-Only NPR

#7

Post by pinjasaur » Thu, Nov 07, 2024, 10:05 PM UTC

zem wrote: post #6

pinjasaur wrote: post #4

Earlier quoted context omitted.

Curious, can you recommend a HTML -> MD converter?

pandoc is pretty good https://pandoc.org/demos.html

D'oh, thanks — I use Pandoc plenty to MD -> HTML I totally spaced it can do the inverse. I was briefly looking at [Turndown](https://github.com/mixmark-io/turndown) for projects in the JS ecosystem.

Re: Text-Only NPR

#8

Post by immibis » Fri, Nov 08, 2024, 12:55 PM UTC

jumploops wrote: post #3

Perfect for LLMs! We use a html->markdown converter for web scraping and sites like this make it even easier/more robust. Side note: bring back the RSS feeds?

Why not just use the HTML ....?