I recently built an end-to-end RAG app, and I needed to do some serious data extraction off the web. I realized that many web scraping tools index too heavily on the markdown style - so you end up with tons of noise in vector DBs and you can't track content diffs.
I built meter to fix that - extracts only the data you want, and notifies you either through webhook or a pull endpoint of any changes to that data over time.