FinCompliance-AI

Project Overview

FinCompliance-AI is an AI-powered RegTech solution designed to automate the detection and reporting of suspicious transactions in full compliance with RBI/PMLA guidelines in India. The system simulates real-time transaction monitoring, applies hybrid risk scoring, and generates FIU-IND compliant Suspicious Transaction Reports (STRs) in XML format with detailed narratives. The project demonstrates a complete end-to-end compliance workflow for banks, NBFCs, and fintechs, reducing manual STR filing time from hours to seconds while ensuring regulatory traceability and explainability.

Live: https://fin-compliance-ai.vercel.app/

Key Objectives Met

• Real-time transaction ingestion and risk assessment
• RBI/FIU-IND compliant STR generation (A–F codes + narrative)
• Live dashboard with risk visualization and alerts
• Integration of mandatory hackathon tools: Cline, Kestra, Oumi, Vercel, CodeRabbit
• Support for domestic & international high-risk scenarios

Features

• Transaction Ingestion: Single, bulk, or live data input with processing simulation
• Risk Scoring:
  • Rule-based (amount, velocity, jurisdiction, PEP, etc.)
  • Simulated XGBoost boost
  • Gemini AI for narrative generation
  • Velocity & location-jump detection
• Dashboard:
  • KPI cards (total txns, alerts, compliance rate)
  • Live volume bar chart & risk pie chart
  • Transaction table with search/filter
  • Velocity alerts (⚡)
• Live Monitoring: Floating Cline agent widget with real-time logs & high-risk alerts
• Workflow Visualization: Kestra-style orchestration steps

Run Locally

1. Clone the repository:

git clone https://github.com/mdkamranalam/FinCompliance-AI.git
cd FinCompliance-AI

2. Install dependencies:

npm install

3. Create .env in the root and ad your Gemini API key:

GEMINI_API_KEY=your-gemini-api-key-here

4. Run the app:

npm run dev

Usage

1. Open the app → see the live dashboard.
2. Use the Transaction Form to add single or bulk transactions (in CSV format).
3. Watch the Kestra workflow animation and Cline agent logs.
4. High-risk transactions appear in red with velocity alerts (⚡).
5. Click any transaction → view risk breakdown, narrative, and XML payload.
6. Download STR XML for FIU-IND demo.

Example Test Case:

Sender Account,Amount,Currency,Receiver Account,Type,Location,Receiver Country
ACC-1001-IN (Rajesh Kumar),980000,INR,CRYPTO-EX-001,RTGS,Delhi,India

License

MIT License — free to use, modify, and share.

Acknowledgments

Built for hackathon using:

• Vercel
• Cline
• Kestra
• Oumi
• CodeRabbit
• Gemini AI