This project is a personal tool for speeding up research process. The main goal is to be able to read/digest papers faster and understand research code (aka messy badly organized code) faster from github. At the moment it is essentially just a paper PDF processor for ChatGPT.
Requires Marker and a sufficient GPU to run it. It has a bit of a lengthy install process but once you have it, it can process PDF files very quickly: Marker
- Add Claude Opus (probably better than ChatGPT for science)
- Github repository parsing
- Better UI/formatting. Primarily so that inline markdown/equations from LLM can be rendered properly.