Sigma Pilot

Website: sigmapilot.ai

Overview

Sigma Pilot is a collective intelligence trading system that learns from the best traders on Hyperliquid, the leading decentralized perpetual futures exchange.

Instead of relying on traditional technical analysis or blindly copy-trading individual wallets, SigmaPilot aggregates wisdom from top-performing traders and generates consensus-based trading signals—only acting when multiple successful traders agree.

"Be as smart as the smartest traders by learning from their collective behavior"

The Problem

Copy-trading a single wallet is risky:

• One trader can have a bad streak
• You're exposed to their mistakes and biases
• No way to know if their edge is real or luck

Traditional signals rely on lagging indicators that don't adapt to changing market conditions.

Our Solution

SigmaPilot solves this by:

1. Monitoring the Best — Continuously tracks 1,000+ traders on Hyperliquid's leaderboard
2. Filtering for Quality — Removes noise traders, HFT bots, and inactive accounts
3. Building a Smart Pool — Selects top 50 qualified traders using Bayesian ranking
4. Detecting Consensus — Generates signals only when multiple independent traders agree
5. Learning Continuously — Updates trader rankings based on realized trade outcomes

Key Features

• Consensus-Based Signals — Only trade when smart money agrees
• Quality-Filtered Alpha Pool — Automatically removes noise and bad actors
• Real-Time Tracking — Live monitoring of top trader positions
• Self-Learning System — Bayesian updates improve rankings over time
• Multi-Gate Validation — 5 independent checks before any signal fires
• Risk-Aware — Filters out low-confidence and high-cost opportunities

How It Works

Leaderboard → Quality Filter → Alpha Pool → Consensus Detection → Signal
   1000+          7 gates         Top 50         5 gates          Trade
  traders                        traders                         signal

Signals only fire when:

• Multiple traders agree on direction
• Traders are statistically independent
• Signal is fresh (not stale)
• Market hasn't moved too far
• Expected value is positive after costs

Getting Started

make install     # Install dependencies
cp .env.example .env
make up          # Start all services
make init        # Initialize Alpha Pool with historical data

Dashboard: http://localhost:4102/dashboard

Run make help for all available commands.

Windows Users: Install make via choco install make or use npm scripts directly. Alternative: npm install && docker compose up -d && npm run init:alpha-pool

Documentation

Document	Description
Architecture	Technical details, services, and APIs
Development Plan	Roadmap and implementation phases
API Reference	REST API and WebSocket documentation
Testing Guide	Test suite and coverage

Technology

Built with TypeScript, Python, PostgreSQL, NATS, and Docker. Event-driven microservices architecture with real-time WebSocket feeds.

Status

Phase 6 In Progress — Multi-exchange integration with Hyperliquid, Aster, and Bybit adapters.

Current capabilities:

• Kelly criterion position sizing with regime adjustments
• Daily drawdown kill switch and circuit breakers
• Market regime detection (trending/ranging/volatile)
• Fail-closed safety posture with retry and metrics
• Multi-exchange interface (Hyperliquid, Aster DEX, Bybit)
• 418 Python + 1,035 TypeScript tests passing

License

PolyForm Noncommercial 1.0.0 — Free for personal and non-commercial use.

For commercial licensing inquiries, please contact us at sigmapilot.ai.