Open source idea

I'm working on a tool that would bundle up informations from pdfs and videos and turn them into beautiful markdown notes for the Obsidian note taking application. At the moment I've developed a simple scraper that uses a pdf reader to scan the pages for tokens, and based on the average words of the page, it takes a screenshot (to be elaborated further) so not to leave image information. It then uses tesseract OCR to scan the images for equations and text, and lastly ot performs a clean of punctuations and useless tokens. I'm dire need for ideas in order to continue the project. I was thinking of implementing chatGPT to automatize some processes by careful prompt engineering.

js, nodejs, GPT, prompt

1