AI RAG DocuQuery

A desktop app for indexing local documents and querying them in natural language through vector search and source-backed answers.

Problem

Finding information across mixed collections of local documents often requires slow, fragmented manual searches.

Solution

The app extracts and indexes content with E5 or MiniLM embeddings and FAISS, retrieves relevant passages from a PyQt6 interface, and can send them to several LLM backends.

Outcome

A local document assistant that exposes passages, files, and source pages and can also work in a citations-only mode without an LLM.

Stack

PythonPyQt6FAISSSentence Transformers