TradingAgents: Multi-agent large language model financial trading framework | Foreign exchange indicator download - MT4/MT5 resources

TradingAgents: Multi-agent large language model financial trading framework https://github.com/TauricResearch/TradingAgents?tab=readme-ov-file 🎉 TradingAgents is officially released! We have received a large number of inquiries about this work and appreciate the community's enthusiastic interest. For this reason we decided to fully open source the framework. We look forward to working with you to create impactful projects! TradingAgents Star History 🚀 Framework introduction | ⚡ Installation and CLI | 🎬 Demonstration video | 📦 Package usage | 🤝 Contribution guide | 📄 Quote TradingAgents framework TradingAgents is a multi-agent trading framework that simulates the operating mode of a real trading company. By deploying professional agents driven by large language models - from fundamental analysts, sentiment analysts, technical analysts, to traders and risk management teams, the platform is able to collaboratively assess market conditions and make trading decisions. The agents also engage in dynamic discussions to determine optimal strategies. The TradingAgents framework is for research use only. Trading performance may vary based on a variety of factors, including the selected underlying language model, model temperature parameters, trading period, data quality and other non-deterministic factors. This framework is not intended as financial, investment or trading advice. Our framework breaks down complex trading tasks into specialized roles. This design ensures that the system adopts a robust and scalable approach to market analysis and decision making. Analyst Team Fundamental Analysts: Evaluate company financial and performance indicators, identify intrinsic value and potential risk signals Sentiment Analysts: Use sentiment scoring algorithms to analyze social media and public sentiment, and judge short-term market sentiment News Analysts: Monitor global news and macroeconomic indicators, interpret the impact of events on market conditions Technical Analysts: Use technical indicators such as MACD and RSI to identify trading patterns and predict price movements The research team is composed of long and short researchers, who will critically evaluate the insights provided by the analyst team. Weigh potential benefits against inherent risks through structured debate. Trader agents integrate reports from analysts and researchers to make informed trading decisions. Determine transaction timing and size based on comprehensive market insights. Risk Management and Portfolio Managers continuously monitor portfolio risk by assessing market volatility, liquidity and other risk factors. The risk management team evaluates and adjusts trading strategies and submits evaluation reports to the portfolio manager for final decision-making. Portfolio managers approve trade proposals. If approved, the order will be sent to the simulated exchange for execution. Installation and Command Line Interface Installation Clone the TradingAgents repository: git clone https://github.com/TauricResearch/TradingAgents.git cd TradingAgents Create a virtual environment using your preferred environment management tool: conda create -n tradingagents python=3.13 conda activate tradingagents Install dependencies: pip install -r requirements.txt Required APIs The FinnHub API is required to obtain financial data, all code is implemented based on the free tier. export FINNHUB_API_KEY=$YOUR_FINNHUB_API_KEY All agents are required to use the OpenAI API. export OPENAI_API_KEY=$YOUR_OPENAI_API_KEY You can run the CLI directly by using the command line: python -m cli.main The interface will display optional parameters: stock code, date, large language model, research depth, etc. The loading results will be displayed in real time when running, and the agent execution progress can be tracked. TradingAgents package implementation details use LangGraph to build TradingAgents to ensure flexibility and modularity. In the experiment, o1-preview and gpt-4o were used as large language models for deep thinking and fast thinking respectively. However, it is recommended to use o4-mini and gpt-4.1-mini during testing to save costs because this framework will initiate a large number of API calls. The Python call imports the tradingagents module in the code and initializes the TradingAgentsGraph() object. The .propagate() function will return the decision result. Main.py can be run, here is a quick example: from tradingagents.graph.trading_graph import TradingAgentsGraph from tradingagents.default_config import DEFAULT_CONFIG ta = TradingAgentsGraph(debug=True, config=DEFAULTCONFIG.copy()) # forward propagate , decision = ta.propagate("NVDA", "2024-05-10") print(decision) You can also adjust the default configuration and set your preferred large language models (LLMs), debate rounds, and other parameters. from tradingagents.graph.trading_graph import TradingAgentsGraph from tradingagents.default_config import DEFAULT_CONFIG # Create a custom config config = DEFAULT_CONFIG.copy() config["deep_think_llm"] = "gpt-4.1-nano" # Use a different model config["quick_think_llm"] = "gpt-4.1-nano" # Use a different model config["max_debate_rounds"] = 1 # Increase debate rounds config["onlinetools"] = True # Use online tools or cached data # Initialize with custom config ta = TradingAgentsGraph(debug=True, config=config) # forward propagate , decision = ta.propagate("NVDA", "2024-05-10") print(decision) For online_tools, we recommend enabling them when experimenting as they provide real-time data access. The agent's offline tools rely on cached data in our Tauric TradingDB - a curated dataset we use for backtesting. We are currently improving the data set and plan to make it publicly available in an upcoming project. Stay tuned! The complete configuration list can be viewed in tradingagents/default_config.py. Get Involved We welcome community contributions! Whether it's fixing bugs, improving documentation, or suggesting new features, your participation will help the project grow. If you are interested in this research direction, please consider joining Tauric Research, our open source financial AI research community. Citation Notes If you think TradingAgents is helpful to you, please cite our work:) @misc{xiao2025tradingagentsmultiagentsllmfinancial, title={TradingAgents: Multi-Agents LLM Financial Trading Framework}, author={Yijia Xiao and Edward Sun and Di Luo and Wei Wang}, year={2025}, eprint={2412.20138}, archivePrefix={arXiv}, primaryClass={q-fin.TR}, url={https://arxiv.org/abs/2412.20138}, }