中文 | English

AITD - AI Trading Agent

🤖 AITD is a local AI trading agent project that can run on almost any machine. You can configure it, validate in paper trading, and run live trading directly in the browser.

Highlights

• Fully customizable trading prompts so you can shape the strategy around your own style
  • Define the AI trader role, core principles, entry rules, and position management rules
  • Save prompts, load historical prompts, and switch to a saved prompt with one click
• Dynamic candidate universe support for flexible symbol selection
  • Useful for sourcing hot symbols from your own workflow
• Custom trading intervals
• Real-time visibility into AI trader inputs and decisions
  • Check whether the input data for each cycle looks correct
  • Check whether the trading actions from each cycle look correct
• Rich portfolio risk controls and detailed trade records
• One-click switching between paper trading and live trading
• Support for major AI model providers
• Fully local runtime so account credentials, prompts, and other sensitive information stay on your machine

Supported Exchanges

1. Crypto:

• ✅ Binance
• OKX (todo)
• BYBIT (todo)
• Gate (todo)

2. Futures:

• CTP (todo)
• IBKR (todo)

Requirements

✅ All you need is Python 3.11+

Quick Start

After downloading the project, run this in the project root:

python3 run.py

You can also specify the port explicitly:

python3 run.py --port 1234

Then open the local address printed in the terminal, for example:

http://127.0.0.1:8788/trader.html

First-Time Setup Suggestions

• Prepare your LLM API_KEY
• Prepare your exchange API Key / Secret, and add your local machine IP to the exchange whitelist if required

1. Start from the Paper Trading page.
2. In AI Model Config, fill in Provider / Model / API Key / Base URL.
3. If you need a proxy, enable it in Proxy Config and enter the proxy address.
4. In the Prompt tab, edit your trading logic and test it to make sure the AI model works correctly.
5. In the Candidate Pool tab, choose one of the following:
   • Static Candidate Pool: manually enter the symbols you want to trade
   • Dynamic Candidate Pool: write a Python function to fetch dynamic symbols
6. Click Start Trading on the current page.
7. Check Trading / Prompt / Log to confirm everything is working as expected.
8. After paper trading runs end to end successfully, configure your live account and start live trading.

Feature Overview

• Trading shows account status, equity curve, open positions, candidate symbols, and recent decisions

• Prompt lets you edit trading logic, adjust position management logic, and test whether the prompt works correctly

• Candidate Pool supports either static symbols or a dynamic Python function

• Log shows runtime logs and error messages from the current process

Key Files

• run.py Startup entry point
• dashboard/ Browser UI assets
• config/ Default configuration files

---

Contact

• X: @ywa_ywa_ywa
• Telegram: @ywa_yl

Copyright and Usage

Copyright © 2026 Yaolin Wang. All rights reserved.

This repository is currently shared for evaluation, learning, and personal reference only. Any commercial use, redistribution, hosted deployment, or integration into paid services requires prior written permission.

For commercial licensing, please contact me through the links above.