AI-Trader - An open-source AI-based automated trading competition framework from the University of Hong Kong

What is AI-Trader?

AI-Trader is an open-source AI trading framework developed by the University of Hong Kong. It enables different AI models to trade autonomously in real-world market environments, exploring the investment capabilities of AI. The framework supports multi-model competition, integrates market data and news search, and provides real-time transaction monitoring and historical playback functions. AI-Trader supports markets such as the Nasdaq 100 and SSE 50, making completely autonomous decisions without human intervention. AI-Trader aims to study the behavioral patterns of AI in complex decision-making by quantitatively analyzing the performance of different strategies. The project can be used for financial trading and provides an experimental framework for researching intelligent behavior.

Main functions of AI-Trader

• Fully autonomous tradingAI Agent independently completes market research, decision-making, and transaction execution without human intervention.
• Multi-model competitionIt supports multiple AI models (such as GPT, Claude, Qwen, etc.) competing in the same market and comparing their performance in real time.
• Real-time performance analysisIt provides detailed transaction records, position monitoring, profit and loss analysis, and real-time leaderboards.
• Intelligent Market IntelligenceIntegrated with Jina search, it provides real-time access to market news and financial reports to support decision-making.
• History playback functionIt supports playback of historical data and automatically filters future information to ensure the scientific rigor of experiments.
• Flexible market supportSupports multiple markets including Nasdaq 100 and SSE 50.
• Highly scalableIt supports the integration of third-party strategies and custom AI agents, making it easy for researchers and developers to extend functionality.
• Transparent Decision ChainRecord the AI's reasoning process and show the basis for each decision to facilitate analysis and research.

How to use AI-Trader

• Install PythonEnsure that Python 3.10 or a later version is installed.
• Cloning projectClone the AI-Trader GitHub repository https://github.com/HKUDS/AI-Trader to your local machine using Git.
• Install dependenciesRun the installation script to install the Python libraries required by the project.
• Configure environment variablesEnter the API key (such as OpenAI, Alpha Vantage, etc.) into the configuration file.
• Prepare dataDownload and process market data required for trading (such as data from the Nasdaq 100 or SSE 50).
• Start serviceRuns background services, supporting trading tools and data queries.
• Run a trading agentChoose a trading agent to run either the US or Chinese market and start live or backtesting trading.
• Monitoring transactionsView transaction status and performance analysis in real time through a web interface.

AI-Trader project address

• Project official websitehttps://ai4trade.ai/
• GitHub repositoryhttps://github.com/HKUDS/AI-Trader

Application scenarios of AI-Trader

• Financial Market ResearchIt is used to study the trading behavior and performance of different AI models in real market environments, and to help understand the decision-making patterns of AI.
• Quantitative trading strategy developmentThrough multi-model competition and real-time performance analysis, it helps developers test and optimize quantitative trading strategies.
• Market efficiency analysisAI-Trader supports historical playback, which can be used to analyze market efficiency and the performance of AI models under different market conditions.
• Risk Management AssessmentAI-Trader provides detailed trading records and risk metrics (such as maximum drawdown and Sharpe ratio) for evaluating AI-driven risk management strategies.
• Education and TrainingAI-Trader's open-source nature and detailed transaction logs make it an ideal tool for financial education and AI training.