Credixa BNPL Platform

Credixa is a comprehensive Buy Now Pay Later (BNPL) platform specifically designed for students, featuring an AI-driven risk engine to assess creditworthiness based on bank statement analysis.

🏗️ Architecture Overview

The platform uses a modern microservices architecture powered by Docker:

1. Frontend (/frontend)

• Tech Stack: React 19, Vite, TailwindCSS, React Router
• Overview: Provides distinct interfaces for Students (to apply for loans, upload documents, and track progress) and Institution Admins (to review applications and monitor metrics).

2. Backend Gateway (/backend-gateway)

• Tech Stack: Node.js, Express, PostgreSQL, JWT, Multer
• Overview: The central API gateway handles user authentication, loan application routing, document uploads, repayment scheduling, and acts as the orchestrator between the frontend and the AI risk engine.

3. AI Risk Engine (/risk-engine)

• Tech Stack: Python, FastAPI, scikit-learn, pdfplumber
• Overview: A dedicated machine learning service that ingests student and parent bank statements (PDFs), extracts behavioral financial features (average balance, overdrafts, risk keywords), and runs a Random Forest model to generate an omniscore (CIBIL equivalent) and a deterministic approval decision.

4. Database (/database)

• Tech Stack: PostgreSQL 16
• Overview: Relational database storing user profiles, loan applications, documents metadata, repayment schedules, and transaction history.

---

🚀 Getting Started

The entire application stack (frontend, backend, AI risk engine, and databases) can be launched seamlessly using Docker Compose.

Running the Services

1. Set up environment variables: Before running the application, copy the .env.example file to .env in the backend-gateway folder:

cp backend-gateway/.env.example backend-gateway/.env

2. Start all services:

docker-compose up --build

This will spin up:

• PostgreSQL database (credixa-postgres on port 5432)
• Redis cache (credixa-redis on port 6379)
• Node.js Gateway (backend on port 3000)
• Python Risk Engine (risk-engine on port 8000)
• React Frontend (frontend on port 5173)

3. Access the Application: The frontend will be accessible at http://localhost:5173.

Database Management Commands

Verify the tables exist:

docker exec -it credixa-postgres psql -U credixa_admin -d credixa_db -c "\dt"

Access the PostgreSQL shell:

docker exec -it credixa-postgres psql -U credixa_admin -d credixa_db

Shut down services and remove volumes (reset database):

docker-compose down -v

---

🔒 Key Features

• Document Vault: Secure upload and storage of academic and financial documents.
• AI-Powered Underwriting: Automated loan decisioning using predictive ML models evaluating non-traditional financial indicators.
• Dynamic Dashboards: Real-time application tracking and repayment schedules.
• Role-Based Access: Dedicated portals and functionality for students and administrators.