An AI-driven paper trading platform built using Node.js and Express, designed to simulate real-world trading decisions using market data, sentiment analysis, and rule-based trade evaluation — without risking real money.

This project focuses on backend architecture, data pipelines, and AI-assisted trade analysis, and is being actively developed as an academic + research-oriented system.

• Introduction
• Key Features
• Project Architecture
• Tech Stack
• Prerequisites
• Installation Guide
• External Dependencies
• Database Setup (Stock Master Data)
• Environment Variables
• Project Status & Roadmap
• Disclaimer
• Author

The AI-Based Paper Trading System is a backend-focused trading simulation platform that evaluates trades using:

• Market data
• News & sentiment signals (planned)
• Rule-based and AI-assisted trade evaluation logic

Unlike real trading platforms, this system operates entirely in paper mode, making it ideal for:

• Strategy testing
• Trade journaling
• Risk classification
• Academic and research use cases

The project follows real-world backend development practices, including modular architecture, version control, and incremental feature development.

• Modular Node.js + Express backend
• MVC-based project structure
• Centralized configuration & utilities
• Live market data ingestion
• Dashboard integration (future scope)
• Trade evaluation logic (WIP)

• News-based sentiment analysis
• AI-assisted trade journal evaluation
• Risk labeling: Good / Risky / Bad trades
• Explainable AI (XAI) for trade decisions

.
├── config/         # App & DB configuration
├── controllers/    # Request handling logic
├── middlewares/    # Custom middleware
├── models/         # Database models
├── routes/         # API routes
├── utils/          # Utility & helper functions
├── public/         # Static assets
├── views/          # Server-rendered views (if used)
├── scripts/        # (Planned) Data insertion scripts
├── app.js          # Application entry point
├── package.json
└── README.md

The architecture is designed to be scalable and extensible, allowing easy integration of AI/ML components.

• Backend: Node.js, Express.js
• Database: MongoDB / SQL (based on implementation)
• APIs: Upstox API (Market data), News API(News data)
• AI/ML (Planned): Sentiment Analysis, Trade Evaluation Models
• Version Control: Git & GitHub

Before installing the project, ensure you have:

• Basic knowledge of JavaScript & Node.js
• A working internet connection
• An Upstox trading account (required for market data)

Download and install Node.js (LTS version recommended):

👉 https://nodejs.org/

Verify installation:

node -v
npm -v

git clone https://github.com/apurvwajage/ai-paper-trading-system.git
cd ai-paper-trading-system

npm install

This installs all required packages defined in package.json.

Create a .env file in the root directory (not included in repo):

PORT=3000
DB_URI=your_database_connection_string
UPSTOX_API_KEY=your_api_key
UPSTOX_API_SECRET=your_api_secret
UPSTOX_REDIRECT_URI=your_redirect_uri

⚠️ Never commit .env files to GitHub.

nodemon app.js

or (for development):

node app.js

This project depends on Upstox for:

• Instrument master data
• Market-related inputs (planned)

You must:

1. Create an Upstox account
2. Generate API credentials
3. Access official documentation

👉 https://upstox.com/developer/api-documentation/

To support trading logic, the system requires stock master data.

From Upstox official documentation:

• Download nse.json (instrument master file)

A script will be added soon to:

• Read nse.json
• Insert all stock entries into the database

📂 Planned location:

scripts/insertNSEStocks.js

Once available, usage will be documented here.

The following environment variables are required:

Current Status: 🚧 Under active development (Work in Progress)

• Complete market data ingestion
• Add NSE stock master insertion script
• Integrate sentiment analysis
• Implement trade risk classification
• Add explainable AI layer
• Deploy backend service

This project is intended only for educational and research purposes.

• No real money trading
• No financial advice
• Paper trading only

Apurva Wajage

Jayesh Varunkar

Srushti Waghmare

Final Year Project – AI-Based Paper Trading System

Feel free to star the repository and follow for updates.