OpenDAN is an open-source Personal AI Operating System designed to consolidate and manage diverse AI modules for personal use. It empowers users to create and orchestrate AI agents, such as assistants and tutors, that can collaborate, integrate with services, and control IoT devices, offering a unified platform for personalized AI experiences.

How It Works

OpenDAN operates as an AI OS, enabling agents to interact via natural language and leverage LLMs, including local models like LLaMA. Agents can be organized into "Workflows" for complex tasks. The system emphasizes a "Personal Knowledge Base" built from user files and emails, accessible by agents, and supports integration with IoT devices and network services through an "AI Environment." A planned "AI Marketplace" aims for one-click installation of AI applications and models.

Quick Start & Requirements

• Install: docker pull paios/aios:latest
• Run: docker run -v /your/local/myai/:/root/myai --name aios -it paios/aios:latest
• Prerequisites: Docker (version > 20.0), OpenAI API Token (optional, for non-local models).
• Setup: Docker installation required. Initial run involves interactive setup.
• Docs: Intro video, Demo, Twitter

Highlighted Details

• Supports local LLMs (e.g., LLaMA) and integrates with Telegram/Email for agent communication.
• Includes built-in agents like a personal assistant (Jarvis), information manager (Mia), and English teacher (Tracy).
• Features an ai_bash tool for developers and a story_maker workflow for AIGC content generation.
• Aims for hardware optimization for local AI model execution and strict privacy management of personal data.

Maintenance & Community

The project is actively developed, with recent merges for MVP (0.5.1) and plans for kernel implementation based on CYFS. Contributions are incentivized via SourceDAO tokens, with a DAO governance page under development.

Licensing & Compatibility

Currently licensed under MIT, with a future transition to SourceDAO. Compatibility for commercial use or closed-source linking is not explicitly detailed beyond the MIT license.

Limitations & Caveats

The project is in its "very early stages" with potential for significant future changes. Support for file formats beyond text and common images is pending. The "OpenDAN Store" is delayed to version 0.5.2.