Neonex Strategy Market

A full-stack Web3 quantitative-trading strategy platform. Users authenticate with an Ethereum wallet (Sign-In with Ethereum), create or fork trading strategies written in Python, backtest them against crypto OHLCV data, run grid-search parameter optimisation, and publish results to a shared marketplace.

Project Layout

backend/    Node.js + Express + TypeScript API (port 3100)
frontend/   Next.js 14 + React 18 UI           (port 3000)
engine/     Python 3.11 + FastAPI compute       (port 3200)

Backend (backend/)

Express REST API that owns authentication, strategy CRUD, versioning, and AI-assisted code generation (via Anthropic Claude). Data is stored in a SQLite database (strategies.db).

Key routes: /api/auth/*, /api/strategies/*, /api/ai/*, /api/symbols/*.

Frontend (frontend/)

Next.js single-page app with TailwindCSS styling, a Monaco code editor for strategy authoring, and Recharts for equity-curve visualisation.

Engine (engine/)

FastAPI service that executes Python backtests using the backtesting.py library, extracts tunable parameters from strategy code, runs grid-search optimisation, and fetches OHLCV data from OKX via CCXT. Supports XLayer token symbols (WOKB, XETH, XBTC, XSOL mapped to their OKX equivalents).

Key routes: /engine/backtest, /engine/validate, /engine/optimize, /engine/scan.

Prerequisites

• Node.js 20+
• Python 3.11+
• npm

Local Setup

1. Backend

cd backend
npm install

# Optional: seed demo data
npx tsx src/seed.ts

# Start dev server (hot-reload)
npm run dev          # listens on http://localhost:3100

Environment variables (defaults work for local dev; see backend/.env.example):

Variable	Default	Description
PORT	3100	HTTP listen port
DB_PATH	./data/strategies.db	SQLite database path
JWT_SECRET	neonex-dev-secret	JWT signing key
PYTHON_ENGINE_URL	http://localhost:3200	Engine service URL
AI_MODEL	claude-sonnet-4-20250514	Model name used by the AI endpoints
OPENAI_BASE_URL	https://cliproxy.exe.xyz/v1	OpenAI-compatible proxy base URL
OPENAI_API_KEY	empty	API key for AI endpoints

2. Engine

cd engine
pip install -r requirements.txt

python3 main.py      # listens on http://localhost:3200

3. Frontend

cd frontend
npm install
npm run dev          # listens on http://localhost:3000

Environment variables (see frontend/.env.example):

Variable	Default	Description
NEXT_PUBLIC_API_BASE	http://localhost:3100	Backend API base
NEXT_PUBLIC_ENGINE_BASE	http://localhost:3200	Engine API base

For local development, point these at http://localhost:3100 and http://localhost:3200 respectively, or update frontend/lib/api.ts.

Architecture

Browser  -->  Frontend (Next.js :3000)
                |
                v
           Backend (Express :3100)  -->  Engine (FastAPI :3200)
                |                              |
           SQLite DB                     CCXT / OKX API

Authentication uses EIP-4361 (SIWE) with JWT sessions. The backend proxies backtest/optimise requests to the Python engine. AI features (strategy generation, improvement, explanation) call the Anthropic Claude API.

Deployment

This codebase was exported from a Railway deployment and then normalized so each service directory can be deployed independently.

Recommended Railway service roots:

• backend/
• frontend/
• engine/

Each service now has a self-contained Docker build, and the frontend also includes a minimal railway.toml. See docs/RAILWAY_DEPLOYMENT.md for details.