oficcejo/aiagents-stock

What it solves

This project provides an automated, multi-agent AI system designed to simulate a professional securities analyst team. It helps retail investors navigate the stock market by automating technical and fundamental analysis, tracking "main force" (institutional) capital flows, and providing data-driven trading signals for A-shares, Hong Kong stocks, and US stocks.

How it works

The system uses a multi-agent architecture where specialized AI agents (e.g., Technical Analyst, Fundamental Analyst, Risk Manager, Sentiment Analyst) collaborate to analyze a stock. It integrates various data sources including AKShare, Tushare, and web-scraping tools (like Playwright for iwencai.com) to gather real-time quotes, financial statements, and macro-economic data from official sources like the National Bureau of Statistics. The system can be deployed via Docker or locally and supports any OpenAI-compatible LLM (such as DeepSeek or GPT-4o) via environment variables.

Who it’s for

It is primarily built for retail investors who want professional-grade analysis without manual research, as well as traders looking for automated screening strategies (e.g., low-valuation value investing, small-cap growth) and those wanting to integrate AI decisions with automated trading via MiniQMT.

Highlights

• Multi-Agent Analyst Team: Simulates a full team of analysts to provide comprehensive reports and a final investment decision.
• Diverse Trading Strategies: Includes pre-built modules for "Main Force" tracking, "Low Price Bull" screening, and macro-cycle analysis (Kondratieff waves and Merrill Lynch Investment Clock).
• Automated Monitoring & Alerts: Real-time price monitoring with entry, take-profit, and stop-loss alerts sent via Email, DingTalk, or Feishu webhooks.
• Quantitative Integration: Supports automated trade execution through MiniQMT integration.
• Batch Analysis: Ability to perform deep AI analysis on dozens of top-performing stocks simultaneously using sequential or parallel processing.