Skip to content
BranchesTags

Repository files navigation

Fenyr AI Trading Agent (Python POC)

⚠️ ARCHIVED - Proof of Concept

This is the Python proof-of-concept version. The production TypeScript version is at: fenyr-ts

Fenyr Logo Python 3.11+ Status

Python Proof-of-Concept for AI Trading Agent

Superseded by fenyr-ts - TypeScript production version

⚠️ This Repository is Archived

This Python version was a proof-of-concept to validate the multi-agent trading architecture. The production version has been rewritten in TypeScript with:

  • Full WEEX SDK integration
  • Type-safe agent tools with Zod
  • Real-time WebSocket support
  • Improved performance

👉 Use the TypeScript version: https://github.com/Miny-Labs/fenyr-ts

📋 What This POC Demonstrated

Fenyr is a production-grade AI trading agent that leverages OpenAI's GPT-5.2 model to autonomously analyze cryptocurrency markets and execute trades on WEEX Exchange. Using advanced function calling capabilities, Fenyr can:

  • Fetch and analyze real-time market data
  • Calculate technical indicators (RSI, EMA, MACD)
  • Make data-driven trading decisions
  • Execute trades with full reasoning transparency
  • Upload AI logs for competition compliance

✨ Features

Feature Description
🧠 GPT-5.2 Integration Latest OpenAI model for superior market analysis
🔧 Function Calling AI directly invokes trading functions
📊 Technical Analysis RSI, EMA, MACD, Bollinger Bands
🔄 Multi-Strategy Momentum, Mean Reversion, Funding Arbitrage
📝 AI Log Compliance Automatic logging for WEEX AI Wars
🛡️ Risk Management Position sizing, stop-loss, take-profit
⚡ Real-time Execution Market orders with sub-second latency

🏗️ Architecture

flowchart TB
    subgraph Input["📥 Data Input"]
        MD[Market Data API]
        TI[Technical Indicators]
        AS[Account Status]
    end

    subgraph Agent["🤖 Fenyr Agent"]
        GPT[GPT-5.2 Engine]
        FC[Function Calling]
        DM[Decision Making]
    end

    subgraph Execution["⚡ Trade Execution"]
        OE[Order Executor]
        RM[Risk Manager]
        AL[AI Log Uploader]
    end

    subgraph WEEX["🌐 WEEX Exchange"]
        API[REST API]
        WS[WebSocket]
    end

    MD --> GPT
    TI --> GPT
    AS --> GPT
    GPT --> FC
    FC --> DM
    DM --> OE
    DM --> RM
    OE --> API
    RM --> OE
    OE --> AL
    AL --> API
Loading

🔄 Trading Flow

sequenceDiagram
    participant U as User/Scheduler
    participant F as Fenyr Agent
    participant G as GPT-5.2
    participant W as WEEX API

    U->>F: Start Analysis
    F->>W: Get Market Data
    W-->>F: Price, Volume, Orderbook
    F->>F: Calculate Indicators
    F->>G: Analyze with Tools
    
    loop Function Calling
        G->>F: Call get_market_data()
        F-->>G: Market Data
        G->>F: Call get_technical_indicators()
        F-->>G: RSI, EMA, MACD
        G->>F: Call get_account_status()
        F-->>G: Balance, Positions
    end

    G->>F: Trade Decision (confidence > 0.7)
    F->>W: Place Order
    W-->>F: Order ID
    F->>W: Upload AI Log
    W-->>F: Success
    F-->>U: Trade Executed
Loading

🚀 Quick Start

Prerequisites

  • Python 3.11+
  • OpenAI API key (GPT-5.2 access)
  • WEEX API credentials

Installation

# Clone the repository
git clone https://github.com/Miny-Labs/fenyr-trading-agent.git
cd fenyr-trading-agent

# Install dependencies
pip install -r requirements.txt

# Configure API keys
cp config.example.py config.py
# Edit config.py with your credentials

Configuration

# config.py
OPENAI_API_KEY = "sk-..."        # Your OpenAI API key
WEEX_API_KEY = "weex_..."        # WEEX competition API key
WEEX_SECRET_KEY = "..."          # WEEX secret key
WEEX_PASSPHRASE = "..."          # WEEX passphrase

GPT_MODEL = "gpt-5.2"            # Model to use
MAX_POSITION_SIZE_BTC = 0.0002   # Max position size
MAX_LEVERAGE = 20                # Competition limit

Running

# Single analysis cycle
python main.py --mode single --symbol cmt_btcusdt

# Demo mode (shows all capabilities)
python main.py --mode demo

# Continuous trading (every 5 minutes)
python main.py --mode continuous --interval 300

📊 Supported Strategies

graph LR
    subgraph Strategies
        RSI[RSI Mean Reversion]
        MOM[Momentum/Trend]
        FUND[Funding Arbitrage]
    end

    RSI --> |"RSI < 30"| BUY[Open Long]
    RSI --> |"RSI > 70"| SELL[Open Short]
    
    MOM --> |"EMA20 > EMA50"| BUY
    MOM --> |"EMA20 < EMA50"| SELL
    
    FUND --> |"Rate > 0.1%"| SHORT[Short Position]
    FUND --> |"Rate < -0.1%"| LONG[Long Position]
Loading
Strategy Signal Action
RSI Mean Reversion RSI < 30 (oversold) Open Long
RSI Mean Reversion RSI > 70 (overbought) Open Short
Momentum EMA20 crosses above EMA50 Open Long
Momentum EMA20 crosses below EMA50 Open Short
Funding Arbitrage Funding > 0.1% Short to collect
Funding Arbitrage Funding < -0.1% Long to collect

📖 API Reference

WeexClient

from weex_client import WeexClient

client = WeexClient(api_key, secret_key, passphrase, base_url)

# Market Data
client.get_ticker("cmt_btcusdt")      # Current price
client.get_depth("cmt_btcusdt")       # Orderbook
client.get_candles("cmt_btcusdt")     # OHLCV data

# Account
client.get_assets()                    # Balance
client.get_positions()                 # Open positions

# Trading
client.place_order(symbol, size, side, order_type)
client.cancel_order(symbol, order_id)

# AI Log (Competition)
client.upload_ai_log(stage, model, input_data, output_data, explanation)

FenyrAgent

from ai_trader import FenyrAgent

agent = FenyrAgent(
    openai_api_key="sk-...",
    weex_client=client,
    model="gpt-5.2",
    max_position_size=0.0002
)

# Single analysis
result = agent.analyze_and_trade("Analyze BTC market conditions")

# Continuous trading
agent.run_continuous(interval_seconds=300)

🔐 AI Log Compliance

Every trade decision is automatically logged to WEEX:

{
    "stage": "Strategy Generation",
    "model": "gpt-5.2",
    "input": {
        "market_data": {"price": 88800, "rsi": 67.65},
        "prompt": "Analyze BTC conditions..."
    },
    "output": {
        "signal": "OPEN_LONG",
        "confidence": 0.74,
        "order_id": "701820686989000728"
    },
    "explanation": "RSI at 67.65 with bullish EMA crossover..."
}

🛡️ Risk Management

flowchart LR
    subgraph Risk["Risk Controls"]
        PS[Position Sizing]
        SL[Stop Loss]
        TP[Take Profit]
        ML[Max Leverage]
    end

    PS --> |"< 2% account"| CHECK{Validate}
    SL --> |"2% below entry"| CHECK
    TP --> |"4% above entry"| CHECK
    ML --> |"≤ 20x"| CHECK
    
    CHECK --> |Pass| TRADE[Execute Trade]
    CHECK --> |Fail| REJECT[Reject Order]
Loading

📁 Project Structure

fenyr-trading-agent/
├── README.md              # This file
├── LICENSE                # MIT License
├── CONTRIBUTING.md        # Contribution guidelines
├── requirements.txt       # Python dependencies
├── config.example.py      # Configuration template
├── config.py              # Your configuration (gitignored)
│
├── main.py                # Entry point
├── ai_trader.py           # Fenyr agent core
├── weex_client.py         # WEEX API client
├── tools.py               # GPT function definitions
│
└── logs/                  # Trading logs (gitignored)

🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

📜 License

This project is licensed under the MIT License - see LICENSE for details.

🏆 Hackathon

Built for AI Wars: WEEX Alpha Awakens

  • Prize Pool: $880,000 USD
  • Competition Period: Dec 2025 - Jan 2026

🔗 Related Projects

Made with 🤖 by Miny Labs

About

GPT-5.2 powered AI trading agent for WEEX AI Wars hackathon

Resources

Readme
MIT license

Contributing

Contributing
Activity
Custom properties

Stars

0 stars

Watchers

0 watching

Forks

1 fork
Report repository

Releases

Packages

Contributors

Languages