A powerful AI black technology that threatens the job security of Silicon Valley fund managers, a popular project in the financial AI field.
Its name is tradingagent, a multi-agent quantitative trading system.
It’s not just a simple trading bot, but rather a collection of AI agents that replicate the workflow of a real trading company.
It includes fundamental analysts, sentiment analysts, news analysts, technical analysts; as well as bullish researchers, bearish researchers, traders, risk management teams, and portfolio managers.
In other words, it doesn’t just give you a “buy/sell” decision on a whim, but simulates a complete investment research meeting.
Its several highlights are obvious:
1. Multi-agent collaboration, not single-point judgment
Traditional AI analysis can easily “decide life or death in one sentence.” TradingAgents breaks tasks into different roles: some analyze financial reports, some review news, some assess market sentiment, some examine technical indicators, and finally, they consolidate into trading opinions.
2. “Bull and bear debate” mechanism
It doesn’t just look for positive signals; it also arranges bullish and bearish researchers to challenge each other. This design is crucial because financial markets fear unilateral narratives; truly valuable analysis often comes from opposing viewpoints.
3. Built-in risk control and portfolio management roles
Trading is not just about judging direction; it also involves considering volatility, liquidity, and risk exposure. TradingAgents includes risk management teams and a Portfolio Manager responsible for evaluating whether trading suggestions should be approved.
4. Supports multiple data dimensions
It can combine fundamental data, news, social sentiment, technical indicators, and more. It’s not just looking at candlestick charts or reading news, but closer to a complete investment research process.
5. Supports multiple large models and local models
Officially supports models from OpenAI, Google, Anthropic, DeepSeek, Qwen, GLM, MiniMax, OpenRouter, Ollama, and others, allowing both cloud-based large models and local models for experimentation.
6. Suitable for research and review, not mindless copying
It also supports decision logs and historical reflection: after each analysis, it records the decision, which can later be reviewed alongside actual returns. This is very valuable for strategy research.
The strongest point of TradingAgents is not telling you which stock will definitely rise, but AI-automating the “investment research decision process.”
It allows ordinary people to see how a trading team gathers information, clashes of viewpoints, risk assessments, and ultimately forms trading suggestions.
Interested users can give it a try:

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

1 Likes