AgenticOS

AgenticOS is an experimental project that explores operating‑system primitives and tooling optimized for multi‑agent AI systems. It treats AI agents as first‑class citizens, focusing on secure resource management, coordinated orchestration, and native AI-to-system integration.

This repository currently provides a developer dashboard and initial UI components for local development and demos. The project is evolving; expect additional services, CLI tools, and architecture modules to be added over time.

Quick project layout

• agentic-ai-os-dashboard/ — Vite + React dashboard used for local development and demos

Quick start — dashboard (local dev)

Prerequisites

• Node.js 18+ (LTS recommended)
• npm 10+ (or compatible package manager)

Steps

1. Change into the dashboard directory:
   • cd agentic-ai-os-dashboard
2. Install dependencies:
   • npm ci (preferred for CI reproducibility) or npm install
3. Configure local environment (optional):
   • Create .env.local and add any keys the dashboard expects (example):
     • GEMINI_API_KEY=your_key_here
4. Start the dev server:
   • npm run dev

The dev server will print a local URL. Default Vite port is 5173, but this project is configured for http://localhost:3000 (see agentic-ai-os-dashboard/vite.config.ts).

Building for production

• From agentic-ai-os-dashboard/ run npm run build. The production output is placed in dist/.

Troubleshooting

• EBADPLATFORM / inotify on macOS

  • Symptom: Unsupported platform errors mentioning inotify.
  • Cause: A Linux‑only dependency was installed or npm install was run in the wrong folder.
  • Fix:
    1. Make sure you run npm ci or npm install from inside agentic-ai-os-dashboard/.
    2. If problems persist: rm -rf node_modules package-lock.json && npm ci.

• Missing package.json at the repository root

  • Only run Node/npm commands inside agentic-ai-os-dashboard/ — the repository root doesn't contain a JS project.

Contributing

• Issues and PRs are welcome. Please open an issue first for larger changes or RFCs.
• Avoid committing secrets — use .env.local (and add it to .gitignore) for local keys.

If you'd like to contribute frontend work, open a PR against agentic-ai-os-dashboard/ and ensure the local dev server builds and runs.

Next steps / roadmap

• Add architecture diagrams and an ARCHITECTURE.md describing core OS primitives for agent lifecycle and resource scheduling
• Add a CONTRIBUTING.md with detailed developer setup and code style guidelines

License

This project is licensed under the terms in LICENSE.