TradingAgents, a revolutionary multi-agent trading framework, has been officially released to the public. This innovative platform mirrors the dynamics of real-world trading firms by deploying specialized LLM-powered agents, including fundamental analysts, sentiment experts, and technical analysts. These agents engage in dynamic discussions to pinpoint the optimal strategy, providing a collaborative approach to evaluating market conditions and informing trading decisions.
The framework is designed for research purposes and supports multiple LLM providers, including OpenAI, Google, Anthropic, xAI, DeepSeek, Qwen, GLM, MiniMax, OpenRouter, and Ollama for local models. Users can import the tradingagents module and initialize a TradingAgentsGraph() object to utilize the framework in their code. The platform also allows for customization of LLMs, debate rounds, and other configuration options.
TradingAgents persists two kinds of state across runs: the decision log and checkpoint resume. The decision log is always on, and each completed run appends its decision to a memory file. The checkpoint resume feature is opt-in and allows the platform to save state after each node, enabling a crashed or interrupted run to resume from the last successful step instead of starting over.
The framework is LLM-driven, and two runs of the same ticker and date can differ due to language model sampling, live data movements, and other factors. To reduce variation, users can lower the sampling temperature or set a non-reasoning model explicitly. Contributions to the framework are welcome, and past contributions are credited per release in the CHANGELOG.md file.
With its modular design and support for multiple LLM providers, TradingAgents provides a flexible and powerful tool for researchers and traders alike. As the field of AI continues to evolve, frameworks like TradingAgents are poised to play a significant role in shaping the future of financial trading and analysis.
Este artículo fue escrito con la asistencia de IA.
News Factory APP - noticias agénticas para impulsar tu SEO y AEO.