Summary PRISM-INSIGHT is a free, open-source AI stock analysis and trading system. It automates stock identification, report generation, and trading simulation/execution, targeting investors and traders seeking advanced, cost-free market insights. The system democratizes access to sophisticated AI financial tools, offering a comprehensive solution from stock detection to strategy simulation.

How It Works A multi-agent architecture with 12 specialized AI agents forms the core. GPT-4.1 powers analysis (technical, financial, market), GPT-5 handles trading simulation, and Claude Sonnet 4.5 manages user interactions. This modular design enables agents to collaborate, producing detailed stock reports, identifying surge stocks, and formulating strategies. It integrates with custom MCP servers for data fetching, web crawling, and search, enhancing analytical depth.

Quick Start & Requirements Requires Python 3.10+, OpenAI/Anthropic API keys, and optionally a Telegram bot token. Installation: clone repo, pip install -r requirements.txt, configure API keys in example files (.env, mcp_agent.config.yaml). wkhtmltopdf is needed for PDF generation. Automated trading requires 한국투자증권 API keys. Linux users may need Korean font installation for charts.

Highlighted Details

• Multi-Agent AI: 12 specialized agents using GPT-4.1, GPT-5, Claude Sonnet 4.5 for layered intelligence.
• Performance Claims: Trading simulator: 408.60% cumulative return (45.1% win rate) in Season 1; 19.47% (66.67% win rate) in ongoing Season 2.
• Integrated Trading: Automated trading via 한국투자증권 API.
• MCP Server Integration: Uses custom MCP servers for stock data, web crawling, search.

Maintenance & Community Discussions via Telegram channel (https://t.me/stock_ai_agent), which also distributes reports. Project gained 100+ stars within six weeks of August 2025 launch, indicating strong initial interest.

Licensing & Compatibility Distributed under the permissive MIT License, allowing broad use, modification, and distribution, including in commercial and closed-source applications.

Limitations & Caveats Lacks real-world trading performance data; metrics are from simulations. Ongoing development (Season 2) suggests potential changes. Reliance on third-party API keys introduces external dependencies and costs. Automated trading is tied to a specific broker's API.