🚀 HiStock AI

"Quantifying Wisdom, Democratizing Alpha."

中文文档 • Quick Start • Features • Architecture

---

HiStock AI is a next-generation Multi-Agent Investment Research & Debate Platform tailored for the A-share market. It's your personal "AI Investment Boardroom" — simulating professional debates between legendary investors (Buffett, Munger, Soros, and more) to help you navigate market noise and uncover hidden risks and opportunities.

---

✨ Features

🎯 Multi-Agent Investment Debate

• AI Boardroom: Multi-round debates moderated by an AI Host with diverse investment personas

• Master Personas: Pre-configured legendary investors including Buffett, Wood, Soros, Lynch, and more

• Context-Aware Dialog: Agents respond to each other, forming a dynamic "think tank"

  

📊 Quantitative Factor Analysis Agent

• Percentile-Based Scoring: Factors scored relative to market distribution (0-100), objectively reflecting stock position

• Weighted Factors: Configurable weights for PE, ROE, Debt Ratio, etc., supporting personalized scoring models

• Multi-Dimension Analysis: Value, Growth, Quality dimensions with customizable weights

• Industry Normalization: (In Development) Relative scoring within industry sectors to eliminate valuation bias

  

🔍 Factor Market

• Real-Time Data: Integrated with AkShare and BaoStock for quotes, valuations, and news

• Factor Analysis: Technical indicators (MA, RSI, MACD) and fundamental metrics (PE/PB, ROE)

• Custom Factors: Create your own quantitative factors using formula DSL

  

🧩 Other Specialized Agents

• News Agent: Tracks real-time market dynamics and extracts key bullish/bearish signals from massive information

  

• Research Agent: Deep dives into company filings and financial reports to provide comprehensive fundamental analysis

  

• Master Agent: Support custom personas, simulating the thinking style and investment logic of legendary investors

  

  

⚙️ Comprehensive Settings

• Model Management: Configure DeepSeek, OpenAI, Google, and more with custom Base URL support

• Agent Configuration: Visually configure agent parameters and prompts

• Appearance: Red/Green up/down color switching to suit different market habits

• Privacy First: All API keys stored locally, with one-click factory reset

  

  

⚡️ Ultimate User Experience

• Streaming Response: Real-time debate stream with progress indicators, making the thinking process visible
• Dark Mode: Carefully designed Dark/Light themes that perfectly adapt to system settings
• Interaction Optimized: Interface and workflows tailored for A-share investors

---

🚀 Quick Start

Prerequisites

• Python 3.12+
• uv (Recommended package manager)
• Node.js 18+ (for frontend)

Installation

# Clone the repository
git clone https://github.com/HiStock-ai/histock.git
cd histock

# Install Python dependencies
uv sync --group dev

# Install frontend dependencies (optional, for development)
cd frontend && npm install && cd ..

Launch

# Mac / Linux
bash start.sh

# Windows (PowerShell)
.\start.ps1

Access the application at http://localhost:1420

Configure LLM

HiStock AI supports multiple LLM providers:

• DeepSeek (Recommended for cost-performance)
• OpenAI (GPT-4o, GPT-4)
• Google (Gemini)
• Anthropic (Claude)

1. Go to Settings (bottom left)
2. Add your API Key

---

🏗️ Architecture

histock/
├── python/                    # Backend (Python)
│   └── histock/
│       ├── agents/            # AI Agents (Master, Quantitative, News)
│       ├── core/              # Core Framework (Agent, Session, Message)
│       ├── services/          # Business Services (Factor, Scoring, Stream)
│       ├── server/            # FastAPI Application & Routes
│       ├── adapters/          # Data Adapters (AkShare, BaoStock)
│       ├── config/            # Configuration Management
│       └── utils/             # Utility Functions
├── frontend/                  # Frontend (React)
│   ├── src/app/               # App Routes & Pages
│   ├── src/components/        # UI Components
│   └── src/lib/               # Frontend Utilities
└── assets/                    # Branding Assets

Tech Stack

Layer	Technology
Backend	FastAPI, Agno Framework, SQLAlchemy
Frontend	React Router 7, Vite, Tailwind CSS, shadcn/ui, Zustand
Data	AkShare, BaoStock, TA-Lib
AI	Multi-Agent Orchestration, Streaming SSE

---

🔮 Roadmap

• Industry Normalization: Relative scoring within industry sectors (SW/CITIC classification)
• Backtesting Module: Historical factor performance analysis
• Portfolio Optimization: Multi-stock portfolio construction
• Mobile App: React Native companion app

---

🙏 Acknowledgements

HiStock AI is proudly built upon the foundation of the ValueCell project. We extend our gratitude to the original creators for their pioneering contributions to the AI investment community.

---

🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

---

⚠️ Disclaimer

HiStock AI is for educational and technical research purposes only. All insights are generated by AI models and do not constitute financial advice. Investing in the stock market involves significant risk.

---

📄 License

This project is licensed under the Apache License 2.0.

---

If you find this project helpful, give us a ⭐