QuantAgent

Multi-agent LLM trading system built on LangGraph, based on the QuantAgent paper.

Four specialized agents analyze OHLC price data in parallel, then a decision agent aggregates their signals into a LONG/SHORT trade with risk parameters. Trades execute on either Deribit Testnet or dYdX v4 Testnet, switchable via EXCHANGE= in .env.

Architecture

START → fetch_data → ┬─ IndicatorAgent ─┐
                      ├─ PatternAgent  ──┤→ RiskDecisionAgent → Execute → Log → END
                      └─ TrendAgent   ───┘

IndicatorAgent — Computes RSI, MACD, ROC, Stochastic, Williams %R locally, then asks Claude to interpret the signals.

PatternAgent — Generates a candlestick chart, sends it to Claude's vision to match against 16 classical patterns (double bottoms, triangles, flags, etc.).

TrendAgent — Fits OLS regression lines to recent highs/lows, generates annotated chart, sends to Claude's vision for trend analysis.

RiskDecisionAgent — Aggregates all three reports, computes ATR-based stop-loss (adapts to volatility per timeframe), sets take-profit via LLM-predicted risk-reward ratio (1.2-1.8x), and schedules a time-based forced exit after the forecast horizon (3 candles) expires.

Setup

# Clone and enter project
cd quantagent

# Create virtual environment
python -m venv venv
source venv/bin/activate  # Linux/Mac
# venv\Scripts\activate   # Windows

# Install dependencies
pip install -r requirements.txt

# Configure
cp .env.example .env
# Edit .env with your API keys

API Keys Needed

1. Anthropic API Key — Get from console.anthropic.com

2. Deribit Testnet (EXCHANGE=deribit) — Sign up at test.deribit.com (no KYC)

   • Go to Account > API Keys > Create new key
   • Enable trade scope
   • Set DERIBIT_TESTNET_API_KEY and DERIBIT_TESTNET_SECRET in .env

3. dYdX v4 Testnet (EXCHANGE=dydx) — v4.testnet.dydx.exchange

   • Connect a wallet and create an account
   • Go to Settings > Export mnemonic (24-word phrase)
   • Get testnet USDC from the faucet (button in the UI)
   • Set DYDX_MNEMONIC and DYDX_ADDRESS in .env

Test your connection:

python test_exchange.py

Usage

# Single analysis, no trade execution (safe first test)
python main.py --once --dry-run

# Single analysis with testnet trade
python main.py --once

# Scheduled hourly execution (BTC only)
python main.py

# Multiple symbols
python main.py --symbols BTCUSDT ETHUSDT

# Different timeframe
python main.py --timeframe 4h --symbols BTCUSDT

Output

Each cycle produces:

• Console summary with decision, entry, SL, TP, justification
• quantagent.log — full execution log
• trade_logs/ — JSON files per trade + running summary

Project Structure

quantagent/
├── main.py              # Entry point + scheduler
├── graph.py             # LangGraph workflow definition
├── state.py             # Shared state schema
├── config.py            # Configuration
├── execution.py         # Exchange execution (Deribit + dYdX)
├── test_exchange.py     # Connection test script
├── agents/
│   ├── indicator.py     # IndicatorAgent
│   ├── pattern.py       # PatternAgent (vision)
│   ├── trend.py         # TrendAgent (vision)
│   └── risk_decision.py # RiskAgent + DecisionAgent
├── utils/
│   ├── data.py          # OHLC data fetching
│   ├── indicators.py    # Technical indicator computation
│   ├── charts.py        # Chart generation (matplotlib)
│   └── llm.py           # Claude API wrapper
├── trade_logs/          # Trade history (auto-created)
├── dashboard/
│   ├── backend/
│   │   ├── app.py           # FastAPI server (port 8001)
│   │   ├── models.py        # Pydantic API models
│   │   ├── trade_analyzer.py # Analytics engine
│   │   └── requirements.txt
│   ├── frontend/
│   │   ├── src/             # React + TypeScript + Tailwind
│   │   ├── package.json
│   │   └── vite.config.ts
│   └── run.sh               # Start both servers
├── requirements.txt
├── .env.example
└── README.md

Dashboard

A React performance dashboard reads from trade_logs/trade_summary.jsonl and visualizes all trading activity.

# Start the backend API (port 8001)
cd dashboard/backend
pip install -r requirements.txt
uvicorn app:app --host 0.0.0.0 --port 8001 --reload

# Start the frontend (in another terminal)
cd dashboard/frontend
npm install
npm run dev
# Opens at http://localhost:5173

# Or start both at once
bash dashboard/run.sh

Pages: Overview (KPIs + equity curve) · Trades (full log table) · Agents (accuracy per agent) · Breakdown (by asset/timeframe/direction) · Settings