Nanobot is an ultra-lightweight, open-source personal AI agent framework written in Python and developed by HKUDS. Instead of a heavy platform, it gives you a compact agent core with a web UI, tool use, memory, and multi-agent workflows that you can self-host on your own machine or server. The project has earned over 49,000 stars on GitHub and was still actively updated in October 2026.
Nanobot is built for people who want a personal AI assistant they fully control. It connects large language models to your chat apps and local tools, so the agent can answer questions, run tasks, and automate routines from the channels you already use. It speaks the Model Context Protocol (MCP), works with OpenAI-compatible APIs and local LLMs, and supports multi-agent setups where several agents collaborate on longer workflows. Everything runs self-hosted, which keeps your data and conversations on infrastructure you own.
Nanobot is completely free and open source under the MIT license. You only pay for any third-party model APIs you choose to connect, or nothing at all if you run it against local models.
Nanobot suits developers and power users who want a personal AI assistant without giving up control of their data. If you like the idea of an agent that lives in your chat apps, remembers context, and automates everyday tasks from your own server, Nanobot is a lean and hackable place to start.