AlphaCouncil  by 164149043

AI-driven multi-agent system for professional stock market analysis and decision-making

Created 3 months ago
292 stars

Top 90.6% on SourcePulse

GitHubView on GitHub
Project Summary

AlphaCouncil AI is a sophisticated, multi-agent system designed for professional A-share market analysis and investment decision-making, leveraging advanced Large Language Models (LLMs). It targets technically savvy users and researchers by simulating a fund company's investment committee workflow, offering AI-driven insights derived from real-time market data. The primary benefit is a structured, LLM-orchestrated approach to complex financial analysis.

How It Works

The system employs a multi-agent architecture comprising 10 AI agents across four hierarchical stages: specialist analysts (macro, industry, technical, fund flow, fundamental), directors (fundamental, market momentum), risk controllers (systemic, portfolio), and a final investment decision-maker. It orchestrates a parallel and serial workflow, integrating real-time A-share data from Juhe Data API with outputs from multiple LLMs (Gemini, DeepSeek, Qwen) via a model-agnostic design. This approach aims to provide a comprehensive, data-driven, and nuanced investment decision process.

Quick Start & Requirements

  • Primary install/run: Deployment via Vercel is the recommended method. Fork the repository and import it into Vercel.
  • Prerequisites:
    • GitHub account for forking.
    • Vercel account.
    • API Keys: Google Gemini, DeepSeek, Juhe Data (for real-time A-share data), and optionally Qwen. These are best configured as Vercel environment variables.
    • Links: Gemini (https://aistudio.google.com/app/apikey), DeepSeek (https://platform.deepseek.com/api_keys), Juhe Data (https://www.juhe.cn/, requires "沪深股票-基本数据" interface), Qwen (https://dashscope.console.aliyun.com/apiKey).
  • Estimated setup time: Vercel deployment is typically quick, minutes to an hour depending on familiarity.

Highlighted Details

  • Agent Roles: Simulates a 10-agent investment committee, covering diverse analytical perspectives from macroeconomics to risk management.
  • Multi-Model Support: Agnostic LLM integration with automatic fallback ensures robustness.
  • Real-time Data: Utilizes Juhe Data API for live A-share market data (5-档盘口, volume, price changes).
  • UI/UX: Features a cyberpunk-themed dark interface with typewriter animations and mechanical keyboard sound effects for an immersive experience.
  • State Persistence: Auto-saves analysis progress (30-minute validity) and history (up to 50 entries) using localStorage.

Maintenance & Community

No specific details regarding maintainers, community channels (like Discord/Slack), sponsorships, or roadmap are provided in the README. The project is attributed to "张一依有把越女剑".

Licensing & Compatibility

The README does not explicitly state a software license. Consequently, its compatibility for commercial use, closed-source linking, or specific distribution terms remains unclear and requires further investigation.

Limitations & Caveats

The system is heavily dependent on third-party API keys, which may incur costs or have usage limits. Saved state data has a short 30-minute validity, and historical analysis may rely on outdated information. A prominent disclaimer warns that all outputs are AI-generated, not financial advice, and may contain "hallucinations" or outdated data, necessitating independent judgment. The absence of a clear license is a significant adoption blocker.

Health Check
Last Commit

1 month ago

Responsiveness

Inactive

Pull Requests (30d)
1
Issues (30d)
0
Star History
95 stars in the last 30 days

Explore Similar Projects

Starred by Chip Huyen Chip Huyen(Author of "AI Engineering", "Designing Machine Learning Systems"), Junyang Lin Junyang Lin(Core Maintainer at Alibaba Qwen), and
4 more.

ai-hedge-fund by virattt

4.7%
48k
AI-powered hedge fund proof-of-concept for educational use
Created 1 year ago
Updated 4 days ago
Feedback? Help us improve.