OP here—thanks for all the replies!
Sounds like the consensus is that scrape data using llm is possible but kinda janky for most cases.
I’ll give Diffbot a shot since a few of you mentioned it. Also like the idea of combining traditional scraping with LLMs for cleanup.
One follow-up: anyone tried this with local LLMs (like Llama 3)? Wondering if it’s any better for privacy-heavy projects.
Sounds like the consensus is that scrape data using llm is possible but kinda janky for most cases.
I’ll give Diffbot a shot since a few of you mentioned it. Also like the idea of combining traditional scraping with LLMs for cleanup.
One follow-up: anyone tried this with local LLMs (like Llama 3)? Wondering if it’s any better for privacy-heavy projects.