RAG chatbot and research agent
Working, in active development
Upload PDFs and ask questions in plain language. The app finds the relevant passages, answers from them, and lists the sources it used, so you can check every claim.
Python, LangChain, Streamlit, Gemini API
What it does today
- Indexes PDFs and retrieves the most relevant chunks for each question.
- Answers with cited sources.
- Runs as a ReAct agent that chooses between PDF search and web search, and remembers the conversation.
What I'm adding next
- A test set to measure and improve answer quality.
- A live deployment with FastAPI and Docker.