Daath AI Parser is an open-source application that uses OpenAI to parse visible text of HTML elements.
https://github.com/kagermanov27/daath-ai-parser
#Open Source AI Parser for HTML Elements
river leaf
hallow tendon
Hi. This seems to use text-davinci-003 for parsing. Is that correct? My app tries to extract useful text from websites but it's really rudimental. At the same time, I think it would be overkill to use davinci for that. Do you have any thoughts on this?
@river leaf
river leaf
@hallow tendon , if you have a project for a specific website that is hard to handle, and your project's prospect is higher than the cost; it is a good way to handle it with a custom parser in text-da-vinci-003 which the Daath AI parser is supporting.
Also if you have a project where maintenance fee (from hiring engineers) is killing the profit, it would come as a good replacement.
My personal use case for it will probably be unit testing for QA of traditional parsers. I already work in a big scraping company, and I see great opportunities in unit testing traditional parsers with precise AI parsers.
Something I forgot, it is also a good source for fine tuning lesser LLM models of all kinds.
hallow musk