nanobot

What is nanobot?

nanobot is an ultra-lightweight personal AI assistant framework developed by HKUDS (Hong Kong University of Data Science) that aims to deliver core AI agent functionality with dramatically reduced complexity. Inspired by OpenClaw, nanobot claims to provide the same essential features with 99% fewer lines of code, making it an attractive option for developers who want to build AI assistants without the overhead of larger frameworks.

Created in February 2026, nanobot has quickly gained traction in the open-source community, accumulating over 32,000 GitHub stars in just over a month. The project is actively maintained with daily updates and has attracted nearly 30 new contributors in its latest release cycle alone.

Getting Started

Installing nanobot is straightforward using Python's package manager:

pip install nanobot-ai

For development or the latest features, you can install directly from GitHub:

pip install git+https://github.com/HKUDS/nanobot.git

nanobot requires Python 3.11 or higher. After installation, you can initialize a new bot configuration:

nanobot init

This creates a basic configuration file where you can set up your preferred LLM provider and chat channels.

Usage & Practical Examples

Basic CLI Usage:

The simplest way to start with nanobot is through its CLI interface:

nanobot chat

This launches an interactive chat session where you can test your AI assistant locally.

Multi-Instance Deployment:

One of nanobot's standout features is its ability to run multiple instances with different configurations:

# Run with custom config
nanobot --config /path/to/custom-config.yaml start

# Run with specific workspace
nanobot --workspace /path/to/workspace start

Platform Integration Example:

Setting up a Telegram bot is remarkably simple. After obtaining a bot token from BotFather, you add it to your configuration:

channels:
  telegram:
    token: "YOUR_BOT_TOKEN"
    allowed_users: ["username1", "username2"]

providers:
  openai:
    api_key: "YOUR_OPENAI_KEY"
    model: "gpt-4"

The framework handles all the complexity of message routing, user authentication, and response formatting automatically.

Performance & Architecture

nanobot's claim of "99% fewer lines of code" than OpenClaw is backed by its minimalist architecture. The core agent functionality is implemented with a focus on essential features rather than comprehensive coverage of edge cases. This approach results in:

• Faster startup times: Minimal dependencies and streamlined initialization
• Lower memory footprint: Efficient resource usage compared to heavier frameworks
• Easier debugging: Smaller codebase makes troubleshooting more manageable
• Rapid development cycles: The project's daily update cadence demonstrates the benefits of a lean architecture

The framework uses modern Python patterns with Pydantic for configuration management, Typer for CLI interfaces, and LiteLLM for provider abstraction, ensuring both performance and maintainability.

Who Should Use nanobot?

nanobot is ideal for:

• Individual developers: Who want to quickly prototype AI assistants without framework overhead
• Small teams: Building personal or internal AI tools with multi-platform requirements
• Researchers: Experimenting with AI agent architectures and need a flexible, lightweight base
• Startups: Requiring rapid deployment of AI assistants across multiple communication channels
• DevOps teams: Looking for simple automation and notification bots

It's less suitable for large enterprises requiring extensive customization, complex workflow orchestration, or guaranteed long-term API stability.

Verdict

nanobot represents an impressive achievement in AI framework design, successfully delivering on its promise of lightweight functionality without sacrificing essential features. The project's rapid growth and active community engagement suggest strong product-market fit for developers seeking simplicity over comprehensiveness. While its alpha status requires caution in production deployments, nanobot's multi-platform support, MCP integration, and clean architecture make it a compelling choice for personal AI assistant projects. For teams prioritizing speed of development and deployment over extensive customization, nanobot offers an excellent balance of functionality and simplicity.

Python Package Manager

pip install nanobot-ai

Development Version

pip install git+https://github.com/HKUDS/nanobot.git

Requirements

Python 3.11 or higher is required.

Initialize Configuration

nanobot init

Start Interactive Chat

nanobot chat

Run with Custom Configuration

nanobot --config /path/to/config.yaml start

Multi-Instance Deployment

nanobot --workspace /path/to/workspace --config custom.yaml start

Basic Configuration Example

channels:
  telegram:
    token: "YOUR_BOT_TOKEN"
    allowed_users: ["username1"]

providers:
  openai:
    api_key: "YOUR_API_KEY"
    model: "gpt-4"