Chat with Multiple PDFs with high accuracy

Now I’m developing AI chatbot based on custom knowledge base.
In other words, it’s Chat with Multiple PDFs.
But PDFs data is very similar so, I’m not sure it is possible to get accurate result.
I have used Langchain and Pinecone vector db.
If you have experience in Chat with Multiple PDFs, please help me.

You might want to look into StanGirard/Guivr on github. It does PDFs, but also goes fairly far beyond that.

Current files it supports:

  • Text
  • Markdown
  • PDF
  • Powerpoint
  • Excel (Not Yet)
  • CSV
  • Word
  • Audio
  • Video

It’s an apache-2.0 license, so useable for commercial purposes, you can include it in patents, etc.

1 Like