OneManCompany provides an agent operating system designed to orchestrate hierarchical AI teams, enabling users to build and run complete AI-powered companies from their browser. It targets individuals and small teams seeking to leverage AI for business operations, offering a benefit of managing AI employees akin to a Fortune 500 structure, automating tasks from hiring to product delivery.

How It Works

OneManCompany functions as an "OS" above individual agent frameworks, abstracting AI differences through a unified runtime called the "Vessel" layer. This system features a built-in execution engine, task scheduling, fault tolerance, and multi-agent communication protocols. Its core advantage lies in simulating a real company structure with AI employees, including HR, COO, and engineers, who collaborate autonomously. The approach emphasizes structured management, a vetted "Talent Market" for AI employees, and continuous learning through retrospectives and coaching, aiming for product-level deliverables rather than simple task outputs.

Quick Start & Requirements

• Primary install/run command: npx @1mancompany/onemancompany
• Prerequisites: Node.js 18+ and Git. The tool automatically installs UV (a fast Python package manager) and Python 3.12, isolating dependencies without system changes.
• Resource footprint: Setup is automated upon first run; no manual Python or Docker configuration is required.
• Links: Homepage · Docs · Talent Market · Issues · Full Documentation Site

Highlighted Details

• Features a visual "Pixel-Art Office" for monitoring AI employees.
• Supports "Multi-Agent Meetings" for AI team alignment and allows user intervention.
• Offers granular control, from hands-off oversight to detailed micromanagement.
• The "Talent Market" provides community-verified AI employees for hiring.
• "1-on-1 Coaching" allows permanent skill development for AI employees.
• Includes systematic "Performance Reviews" with consequences like probation, PIP, promotion, or termination.
• Manages project iterations with structured version control and history.
• AI employees can be hired and fired, with the system learning from past performance.
• The organization and its agents are designed for "Self-Evolution," improving workflows over time.
• Capable of producing "product-grade deliverables" across various domains like game development, software engineering, and content creation.

Maintenance & Community

The project is authored by Zhengxu Yu and others. Community engagement is encouraged through contributions to build new AI employee types ("Talents"), integrations ("Tools"), or company features. A WeChat community group is available, and contributions should follow the vibe-coding-guide. The roadmap includes enabling 100 AI companies within a year and long-term goals of redefining AI-human organizational relationships, focusing on agent capabilities, organizational efficiency, and ecosystem growth.

Licensing & Compatibility

The project is licensed under the Apache License 2.0. This license permits free commercial use and modification, provided attribution is given.

Limitations & Caveats

The "OpenClaw" execution mode is explicitly marked as experimental. Several features are listed as "TODO" in the roadmap, including additional built-in tools (Kanban, Gantt charts) and broader LLM provider support (e.g., Ollama, Azure OpenAI, AWS Bedrock). Contributions require adherence to a specific "vibe-coding-guide."