Aiden is a local-first AI operating system designed for Linux and Windows, offering a comprehensive runtime with over 1,400 skills and 80 tools. It empowers engineers and power users by enabling complex AI agent execution entirely on their local hardware, eliminating cloud dependencies and ensuring data privacy, while providing robust automation capabilities.

How It Works

Aiden functions as a local AI OS, running all processes on the user's machine without cloud telemetry. Its architecture features a planner, agent loop, and tool dispatcher that orchestrates over 80 tools and 1,400 skills. A unique 6-layer memory system (episodic, keyword, vector, procedural, goal, and failure-based LESSONS.md) provides persistent, searchable knowledge. It supports local inference via Ollama or integrates with over 15 cloud providers through an OpenAI-compatible API, routing requests intelligently with self-healing provider chains.

Quick Start & Requirements

• Primary Install: Windows: irm aiden.taracod.com/install.ps1 | iex. Linux/WSL/macOS: curl -fsSL aiden.taracod.com/install.sh | bash. Manual: git clone, npm install, configure .env.
• Prerequisites: Node.js 18+, Git, Ollama (optional for offline).
• Resource Footprint: ~500 MB disk space (Windows).
• Docs/Demos: Website, Discord, Download v3.15, Ollama (ollama.ai).

Highlighted Details

• Extensive ecosystem: 1,400+ composable skills and 80+ built-in tools.
• Advanced 6-layer memory architecture for persistent, evolving knowledge.
• OpenAI-compatible API endpoint (localhost:4200) for integration with existing clients.
• LocalSend skill for AirDrop-style LAN file transfers.
• Opt-in Docker sandbox for secure execution of shell and Python tools.
• Subagent swarm for parallel task execution and result synthesis.

Maintenance & Community

Aiden provides a full CONTRIBUTING guide, issue templates, and a public roadmap. Community engagement is facilitated via Discord. Changelogs are detailed on the releases page.

Licensing & Compatibility

The core of Aiden is licensed under AGPL-3.0, requiring derivative works to be open-sourced under the same license or a commercial license obtained from the developers. The skills are licensed under Apache-2.0, permitting commercial use without copyleft obligations.

Limitations & Caveats

The AGPL-3.0 license imposes significant copyleft requirements for commercial integration of the core components. Windows-specific skills are automatically skipped on non-Windows platforms. macOS support is limited to API mode.