Feature Requests

I recently paid for BoltAI (and Mindmac) and quite frankly BoltAI is an amazing piece of software in the way the plug ins now work. In particular the way one can connect claude/openAI to say perplexity (search) or brave (search) is an incredibly powerful tool for doing scientific research. The big missing piece here is having a RAG/PDF reader processor that is done in the way that PDFPals is able to highlight selective parts of the text and summarize it. It would be really great if one could merge these two pieces of software in someway as a plugin and eventually include a RAG component (e.g. being able to summarize / RAG say an entire Zotero library) as well. Right now PDFPals feels a bit orphaned (Claude 3 vs Sonnet 3.5 , no ability to search the web for additional info etc ) if you could glue these together (like mindmacs PDF reader) that would be awesome. Regardless really nice work on both pieces of software to be honest!

It would be great if alongside to the chatbot there was the possibility to highlight and add notes to the PDF itself.

Sometimes I just want to OCR my pdf and use it through my default pdf viewer. Can you add a feature in PDF Pal to just OCR it using the open source or text-embedding OpenAI model and embed the OCR results into the original pdf for external use, or generate a new pdf with it.

The app is the fastest I've seen so far. However, I noticed it doesn't use more than 4 to 5 chunks. It would be great to have an option to configure this setting and add more chunks, even if it increases the price per prompt.

As mentioned on twitter a while back: I suggest the development of an interactive form filling feature for PDF Pals. This feature would involve an intelligent agent asking users specific questions related to form fields, and potentially even filling them out on behalf of the user. This could significantly simplify the process of completing PDF forms, making it more user-friendly and efficient. By transforming the form filling experience into an interactive dialogue, we can make PDF Pals an even more valuable tool for users working with PDFs.

somehow AI still make mistakes while reading a whole book, its' best to only to summarize what's highlighted

I have a 2017 MacBook Air. Would love if you could support the this.