Regulatory RAG — AI-Powered Financial Compliance Assistant

Local Retrieval-Augmented Generation (RAG) system built with FastAPI, Ollama (Llama 3.1 8B), and ChromaDB to answer Basel III, OCC, and FFIEC regulatory questions with cited responses.
Includes document ingestion, semantic search, and regulator-aware retrieval.

Quickstart

# 1) create venv
python3 -m venv .venv && source .venv/bin/activate

# 2) install
pip install -r requirements.txt

# 3) (optional) set local LLM backend via Ollama
ollama serve
ollama pull llama3.1:8b

# 4) ingest your PDFs into data/source_docs/{basel,ffiec,occ}/ then:
make ingest

# 5) run API
uvicorn app.api.main:app --host 0.0.0.0 --port 8000