A fully self-contained, real-time simulated crypto and stock exchange built for NextBull x IIT Kharagpur OpenSoft 2026. No external market data. No price feeds. Every price, every trade, every candlestick is generated locally.

All three start with a single command. The bots and terminal wait for the engine's health check before connecting.

git clone <repo>
cd opensoft
docker-compose up --build

That is it. No environment setup required. Default secrets are baked into the compose file for local use — change them before any public deployment.

# Stop all services
docker-compose down

# Stop and wipe the database volume (full reset)
docker-compose down -v

# Rebuild a single service after code changes
docker-compose up --build exchange-engine
docker-compose up --build web-terminal
docker-compose up --build trading-bots

Each service has its own README with local dev instructions. Short version:

# Exchange engine
cd exchange-engine
go run ./cmd/server

# Web terminal
cd terminal
npm install
npm run dev

# Trading bots
cd trading-bots
pip install -r requirements.txt
python main.py

The engine must be running before starting the bots or the terminal.

All are set in docker-compose.yml. For local dev, copy .env.example inside each subdirectory.

Go to http://localhost:8080/admin, enter the admin secret (default: admin).

From there you can:

• View all registered users and their portfolio values
• Seed a demo account with 30 days of realistic history (trades, equity curve, calendar entries, journal notes) — useful for showing the dashboard without having to trade manually for a month
• Reset any user's portfolio back to $100,000

The admin panel is compiled into the engine binary (//go:embed admin.html) so it is available wherever the engine is deployed.

There is also a standalone Python seed script at the project root:

pip install requests
python seed_demo.py

This registers a user (username: user, password: user), places ~80 orders across all symbols, and writes 30 days of portfolio snapshots and 15 notes directly to SQLite. Useful for seeding a local dev database without going through the admin panel.

                        REST / WebSocket
exchange-engine:8080 <─────────────────────> terminal:3000
        ^
        │  REST  (same public API as human traders)
        │
trading-bots:9090 (metrics only, no inbound connections from engine)

Persistence: SQLite WAL mode, mounted as Docker volume "exchange_data"

The engine has no knowledge of the frontend or the bots. Both connect as regular API clients using JWT tokens. The bots register accounts on startup and appear on the leaderboard like any other user.

Prices are simulated using Geometric Brownian Motion. They diverge from real market prices immediately after startup — that is by design.

opensoft/
├── docker-compose.yml
├── exchange-engine/       Go backend
├── terminal/              Next.js frontend
├── trading-bots/          Python bots