TradingAgents-astock
A股多Agent投研框架 — 适配A股数据源(龙虎榜/游资/解禁等)，7位分析师基于A股规则的辩论决策，基于TradingAgents深度改造，适配大A。A-share multi-agent investment research framework — 7 AI analysts, bull/bear debate, risk assessment。
TradingAgents-Astock is a Python framework that uses multiple AI agents working together to analyze stocks listed on China's A-share market — a fork of an existing multi-agent trading research tool that was originally designed for US stocks. The problem it addresses is that Chinese A-share markets have unique rules, data sources, and market dynamics (such as daily price limits, a T+1 rule where you can't sell on the same day you buy, and the influence of hot-money traders and government policy) that US-focused tools don't account for.
The framework works by running seven specialized AI analyst agents in sequence, each examining a different angle: market technicals, social sentiment, news, company fundamentals, government policy, hot-money flow, and share lockup expiry events. Their reports feed into a bull-versus-bear debate between two researcher agents, followed by a risk assessment from three perspectives (aggressive, conservative, neutral), and finally a portfolio manager agent that outputs a Buy/Hold/Sell recommendation. All data comes from free Chinese financial data sources with no API keys required.
You would use this if you're researching Chinese A-share stocks and want a structured, AI-driven analysis that understands local market rules rather than generic stock-market logic. It includes both a command-line interface and a web-based interface for running analyses without writing code. Reports are output in Chinese by default.