FinCore AI is a unified AI-powered financial platform designed to solve four critical challenges in Nigeria’s financial ecosystem:
- Fragmented banking data
- Poor cash flow visibility
- Limited access to credit
- Rising fraud threats
Built for the GTCo Squad Hackathon 3.0, the platform leverages Squad APIs, Open Banking, and Machine Learning to deliver real-time financial intelligence.
- Unified dashboard across all accounts
- User-defined account labeling (Personal, Business, Savings, etc.)
- Real-time transaction synchronization
- AI-powered forecasting (30/60/90 days)
- Virtual accounts via Squad
- Automated reconciliation via webhooks
- Alternative credit scoring using behavioral data
- Random Forest model
- Instant loan disbursement via Squad
- Variational Autoencoder (VAE) anomaly detection
- Graph-based fraud pattern detection
- Real-time transaction blocking
Frontend:
- React / React Native
Backend:
- Node.js / FastAPI
Machine Learning:
- TensorFlow / PyTorch / Scikit-learn
Infrastructure:
- Docker / Kubernetes
Integrations:
- Squad API (payments, transfers, accounts)
- Mono (open banking)
- KYC Providers (BVN/NIN verification)
See:
- docs/architecture/system_diagram.png
- docs/architecture/data_flow.png
- Squad APIs (Virtual Accounts, Transfers, Subscriptions)
- Mono Connect SDK
- KYC Verification APIs
Postman Collection: docs/api/squad_postman_collection.json
| Model | Purpose |
|---|---|
| ARIMA / LSTM | Cash flow forecasting |
| Random Forest | Credit scoring |
| VAE | Fraud anomaly detection |
| GNN | Fraud network analysis |
Model cards available in: docs/ml/model_cards/
git clone https://github.com/your-team/fincore-ai
cd fincore-ai
cp .env.example .env
docker-compose up- User onboarding (BVN/NIN)
- Bank account linking (Mono)
- Dashboard aggregation
- Cash flow prediction
- Credit scoring + loan disbursement
- Fraud detection in real time
- PRD → docs/PRD/
- Architecture → docs/architecture/
- ML Models → docs/ml/
- API Specs → docs/api/
Built for GTCo Squad Hackathon 3.0
MIT License