PDFConverse

Chat with your PDF!

Upload your PDF and talk to it in natural language to gain insights and knowledge, just as you would talk to an LLM such as ChatGPT or Bard.

How it works
  • PDF is divided up into small chunks
  • Chunks are embedded using an embedding model
  • Embeddings are stored in a vector store
  • User asks a question
  • Question is embedded using the same embedding model
  • Similarity search of the embedded question is performed with docs in the vector store
  • Question + similar docs are sent to LLM
  • LLM answers the question, which is shown to the user
Demo
Project: https://github.com/piyush-jaiswal/PDFConverse