FinLLM is an open-source assistant for financial document understanding and compliance automation, powered by OCR + LLMs + Vector Search (RAG). It enables users to extract content from scanned or digital documents and chat with them contextually.
- ✅ Upload scanned PDFs, DOCX, or image files
- ✅ Perform OCR using Tesseract
- ✅ Chunk and embed content using sentence-transformers or OpenAI
- ✅ Store in FAISS vector database
- ✅ Run RAG-style querying using OpenAI or DeepSeek
- ✅ CLI and Streamlit-based chat interface
- ✅ Easily extensible for other domains (legal, healthcare, HR, etc.)
pip install git+https://github.com/akash049/fin-llm-assistant.git