DocChat AI — PDF Question Answering with Citations
A retrieval-augmented document assistant that turns uploaded PDFs into searchable knowledge bases and returns grounded answers with page citations.
Status:Deployed portfolio application
View source on GitHub (opens in a new tab)- Services
- AI Automation, Machine Learning
- Stack
- Python, LangChain, FAISS, Groq +3 more
01
Our solution
The application extracts and chunks PDF text, creates local MiniLM embeddings, indexes content in FAISS, retrieves relevant passages, and uses Llama 3.3 through Groq to generate cited answers.
02
The outcome
Reduces time spent manually scanning long PDFs and demonstrates a practical trust layer for document-based AI through source citations.
03
Best for
Legal and policy search, HR knowledge bases, product manuals, support enablement, and compliance document review.
Technology
The stack behind it
- Backend
- Python
- AI & ML
- LangChain
- FAISS
- Groq
- Llama 3.3
- MiniLM
- Frontend
- Streamlit