📘 ClaimRAG-Engine

AI‑Powered Insurance Claims Processing & Intelligence System with Streamlit Chatbot UI

ClaimRAG-Engine is an AI-native claims automation and intelligence platform built with:

• FastAPI backend for RAG, LLM agents, workflows
• Streamlit frontend for interactive chatbot & dashboards
• RAG ingestion pipeline
• Fraud detection intelligence
• Human-in-the-loop review flow

The system enables policy interpretation, claim evaluation, document analysis, and real-time Q/A using modern Retrieval-Augmented Generation.

---

✨ Features Overview

🔍 RAG + LLM Agent Capabilities

• Context-aware policy lookup
• Claim coverage interpretation
• Document-based answer generation
• Multi-step reasoning engines
• Custom agent workflows

🤖 Automated Claim Processing

• Eligibility validation
• Coverage constraints extraction
• Missing info detection
• Automated decision support

🔒 Fraud Intelligence

• ML + LLM signal verification
• Risk score generation
• Suspicious activity detection
• Manual review triggers

📄 Document Workflow Automation

• Policy ingestion
• PDF → text parsing
• Cleanup, chunking & vectorization
• Vector store indexing

💬 Streamlit Frontend

• Chatbot interface
• Real-time RAG responses
• Claim submission UI
• Logs / analytics (optional)

---

🏗 Project Structure

ClaimRAG-Engine/
│
├── backend/
│   ├── src/
│   │   ├── api/
│   │   ├── core/
│   │   ├── ingestion/
│   │   ├── rag/
│   │   └── agents/
│   ├── main.py
│   ├── pyproject.toml
│   └── uv.lock
│
├── frontend/
│   ├── app.py
│   ├── components/
│   ├── utils/
│   └── requirements.txt
│
├── .gitignore
├── README.md
└── LICENSE (optional)

---

⚙️ Backend Setup (FastAPI)

1. Move into backend

cd backend

2. Create environment

uv venv
.\.venv\Scripts\activate

3. Install dependencies

uv sync

4. Run API server

uvicorn backend.main:app --reload

API will run at:

📍 http://localhost:8000
📍 http://127.0.0.1:8000/api/v1/query/ (Swagger UI)

---

📥 RAG Document Ingestion

To ingest policy documents:

uv run -m src.ingestion.ingestion

This will:

• Load PDFs / docs
• Clean + chunk text
• Create embeddings
• Store into vector database

---

💬 Frontend Setup (Streamlit UI)

1. Move into frontend

cd frontend

2. Install UI dependencies

pip install -r requirements.txt

3. Run streamlit

streamlit run app.py

Streamlit UI will run at:

📍 http://localhost:8501

---

🔌 Frontend–Backend Communication

The Streamlit UI communicates with the FastAPI backend through endpoints like:

POST /api/v1/query

For RAG-based chatbot responses.

POST /api/v1/claims/evaluate

For claim scoring & eligibility.

POST /api/v1/fraud/check

For fraud detection workflows.

Make sure the FastAPI backend is running before starting the Streamlit UI.

---

📘 Environment Variables

Create backend/.env:

OPENAI_API_KEY=your_key_here
VECTOR_DB_PATH=./data/vector_store
MODEL_NAME=gpt-4o-mini

Create frontend/.env:

API_URL=http://localhost:8000

⚠️ Never commit .env to Git.
Add to .gitignore.

---

🚀 Run the System End-to-End

1️⃣ Start backend

cd backend
uvicorn backend.main:app --reload

2️⃣ Start frontend

cd frontend
streamlit run app.py

3️⃣ Interact with the chatbot UI at:

http://localhost:8501

---

🧱 Tech Stack

Component	Technology
Backend	FastAPI, LangChain, RAG
Frontend	Streamlit
Embeddings	Gemini
Vector Store	Postgres-Pgvector
Agents	LangChain Agents
Packaging	UV

---

🌟 Support

If this project helps you, consider starring ⭐ the repo!

---