OpenDAN is an open-source Personal AI OS: a self-hosted system that consolidates many AI modules into one personal environment you control. Short for “Open and Do Anything Now”, it combines a voice-and-chat assistant core with pluggable AI apps, long-term memory and tool integrations, aiming to be a private, always-on AI layer for an individual rather than another cloud chatbot. The project has around 2,100 GitHub stars and remains in active development, with code pushed in 2026.
OpenDAN runs on your own hardware (a home server or spare PC) and presents a unified assistant that can take voice input, manage personal knowledge, and call external services and models. Its modular design lets you add capabilities such as news digesting, note capture and smart-home style control through AI apps, while personal data stays on your machine. You can connect it to OpenAI-class models for reasoning while keeping the operating layer, memory and orchestration under your own control.
Free. OpenDAN is open source under the MIT license; you pay only for your own hardware and any third-party model API usage.
Privacy-conscious tinkerers and homelab users who want a personal AI assistant they fully own, and developers interested in modular, local-first agent architectures.
OpenDAN’s roadmap centers on making the personal AI OS feel like a real operating layer: a home screen of AI apps, a voice-first interaction model, and connectors that let the assistant act across your files, feeds and devices. Because it is self-hosted, it appeals to users who want the convenience of an always-on assistant without sending their personal context to a third-party cloud, and developers can write new AI apps against its module system.