Nanobot Developer Guide
Nanobot is a lightweight open-source AI agent for tools, chats, and workflows. This guide is the practical developer entry point: what Nanobot is, how to get it running, how the runtime is shaped, where configuration lives, and where to start if you want to extend it.
This guide is based on the official Nanobot docs and the official HKUDS/nanobot repository, checked on June 26, 2026.
In this section​
- Setup and first run
- Architecture and runtime
- Configuration and security
- Extending and contributing
- Simple user guide
1. What Nanobot is​
The shortest accurate description is:
Nanobot is a Python-based agent runtime that combines LLM access, tool execution, chat-style interaction, and workflow automation in a relatively small package.
That is important because Nanobot is not only a CLI toy and not only a web chat frontend. The upstream docs and repo position it as a practical agent you can install locally, wire to your providers and tools, and use for both interactive and automated work.
2. The mental model​
Think of Nanobot as four cooperating layers:
| Layer | What it does |
|---|---|
nanobot core | Runs the agent loop and tool orchestration |
| bridge and connectors | Connect external surfaces and services |
| web UI | Gives you a browser-facing control surface |
| config and scripts | Handle install, startup, and runtime wiring |
This makes Nanobot attractive when you want something more structured than a one-file script, but less platform-heavy than a full multi-service agent stack.
3. Why developers would choose it​
- It stays lightweight compared with larger agent platforms.
- It uses a Python-first stack, which lowers the barrier for many teams.
- It is aimed at tools, chats, and workflows, not only code generation.
- It has an official architecture and development guide, so the project is easier to study than many small agent repos.
4. Where to look first​
If you are new to the project, the most useful order is:
- the Nanobot overview page,
- the install and quickstart flow,
- the architecture notes,
- the development and contribution docs,
- the repo structure once you know what you are looking for.