Summary

Luopan is a research system designed for quick, judgment-oriented analysis of industries and companies. It targets technical users needing to assess investment or career opportunities by clarifying money flow, power dynamics, and strategic positioning, moving beyond mere data aggregation.

How It Works

Luopan routes research requests to industry or company-specific modules. Its methodology combines Source Grading (A/B/C), a Multi-Source Perspective Matrix, Adversarial Verification, and an Honesty Principle. Industry analysis covers structure, power, profits, and opportunities. Company research tailors insights for investment (financials, moat, valuation) or job seeking (prospects, stability, career capital), preventing user goal conflicts.

Quick Start & Requirements

Clone the repository: git clone https://github.com/zhangxiaoqiang1991/luopan.git. Requires preserving the full directory structure. Integrates with various AI assistants (e.g., Claude Code, Codex, KimiCode).

Highlighted Details

• Delivers "judgments" focused on "money and power," not encyclopedic data.
• Differentiates analysis for distinct user goals: investment vs. job seeking.
• Employs structured methodology: source grading, multi-source perspectives, adversarial verification.
• Provides "Core Judgments" with data, sources, and "So what?" analysis.
• Analyzes "power/discourse power" within value chains.
• Features detailed case studies for prominent companies and industries.

Maintenance & Community

Developed by "大厂转型人强哥" (Zhangxiaoqiang), a former Tencent/ByteDance employee focused on AI applications. Community interaction is primarily via WeChat (paid groups/consulting) and Twitter (@andyxqzhang001).

Licensing & Compatibility

Released under the permissive MIT License, allowing for commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

Analysis prioritizes judgment over exhaustive data recall. For unlisted companies, "limited public information" is clearly marked, avoiding fabricated data from rumors. Job-related analysis guides interview decisions, not final offer acceptance. Investment reports strictly separate company quality from price, requiring explicit valuation assumptions and safety margins. Vague sources are avoided, and fabricated data is not used.