⚡ Nexus Trade Engine

AI-native plugin trading framework with full cost modeling.

Nexus is a modular, plugin-driven algorithmic trading platform designed for steady portfolio growth. It treats transaction costs, taxes, slippage, and spread as first-class citizens — not afterthoughts — ensuring strategies that backtest well also perform in production.

---

Architecture

Nexus is built on a five-layer architecture where every component is independently scalable and replaceable:

Layer	Purpose	Key Components
Presentation	Web dashboard & API consumers	React UI, WebSocket streams
API Gateway	Orchestration & auth	FastAPI, JWT/RBAC, rate limiting
Core Engine	Trade execution & risk	Order manager, cost model, risk engine
Plugin System	Strategy marketplace	SDK, registry, sandboxed runtime
Data Layer	Storage & market feeds	TimescaleDB, PostgreSQL, Redis

Interactive architecture diagrams are available in docs/architecture/.

Core Concepts

Plugin-First Strategy System

Strategies are self-contained plugins that implement the IStrategy interface. Developers have complete freedom in their implementation — fixed algorithms, neural networks, LLM calls, or any hybrid combination. The engine only cares about the signals that come out.

from nexus_sdk import IStrategy, Signal

class MyStrategy(IStrategy):
    def evaluate(self, portfolio, market, costs):
        # Your logic here — anything goes
        return [Signal.buy("AAPL", weight=0.7)]

Three Execution Modes, One Interface

Every strategy runs identically across all three modes:

• Backtest — Historical simulation with full cost model
• Paper Trade — Live market data, simulated execution
• Live Trade — Real money, real broker, same interface

Cost Model as Input, Not Afterthought

The ICostModel is passed directly into every strategy's evaluate() call. Strategies can (and should) factor in commissions, spread, slippage, taxes (FIFO/LIFO), wash sale rules, and dividend withholding before emitting signals.

Quick Start

Prerequisites

• Docker & Docker Compose
• Python 3.11+
• Node.js 20+
• Git

Setup

# Clone the repository
git clone https://github.com/your-org/nexus-trade-engine.git
cd nexus-trade-engine

# Start infrastructure (Postgres, TimescaleDB, Redis)
docker compose up -d

# Install engine dependencies
cd engine
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

# Run database migrations
python -m alembic upgrade head

# Seed sample market data
python ../scripts/seed_data.py

# Start the engine
uvicorn main:app --reload --port 8000

# In a separate terminal — start the frontend
cd frontend
npm install
npm run dev

Install the SDK (for strategy developers)

pip install nexus-trade-sdk
# or from source:
cd sdk && pip install -e .

Project Structure

nexus-trade-engine/
├── engine/                 # Core trading engine (FastAPI)
│   ├── core/               # Order management, portfolio, risk
│   │   └── execution/      # Backtest / Paper / Live backends
│   ├── plugins/            # Plugin SDK, registry, sandbox
│   ├── data/               # Market data feeds & providers
│   ├── events/             # Pub/sub event bus
│   ├── api/                # REST & WebSocket routes
│   └── db/                 # Models, migrations, session
├── sdk/                    # Installable SDK for strategy devs
│   └── nexus_sdk/          # IStrategy, Signal, types, testing
├── strategies/             # Example strategy plugins
│   └── examples/           # Reference implementations
├── frontend/               # React dashboard
├── tests/                  # Test suite
├── scripts/                # DB init, data seeding, utilities
├── docs/                   # Architecture docs & diagrams
└── docker-compose.yml      # Infrastructure stack

Developing a Strategy Plugin

1. Install the SDK: pip install nexus-trade-sdk
2. Create a manifest file (strategy.manifest.yaml)
3. Implement the IStrategy interface
4. Test locally with the backtest runner
5. Publish to the marketplace

See the Plugin Developer Guide for full documentation.

Tech Stack

Component	Technology
Engine	Python 3.11, FastAPI, Celery
Database	PostgreSQL 16, TimescaleDB
Cache	Redis 7
Frontend	React 18, Vite, Tailwind CSS
Task Queue	Celery + Redis broker
Containerization	Docker, Docker Compose
Testing	pytest, pytest-asyncio

Roadmap

• Core architecture scaffold
• Plugin SDK v1.0
• Backtest engine with full cost model
• Paper trading with live data feeds
• React dashboard MVP
• Strategy marketplace
• Live broker integration (Alpaca, IBKR)
• Multi-asset support (crypto, forex, options)

License

MIT License — see LICENSE for details.