Tired of opaque, heavy AI platforms that take control away from you? Meet Nanobot, an ultra-lightweight personal AI agent designed for true ownership and local control.
Most existing frameworks feel bloated and difficult to manage, making it hard to maintain privacy or customize the core logic. Whether you are dealing with complex tool calls like those in smolagents or managing massive infrastructure, you need a lightweight alternative.
The Problem: The Complexity of Modern AI Agents
As AI evolves, we are seeing a surge in complex orchestration layers. While powerful, these systems often create a barrier to entry for developers who want a portable and readable agent core.
Managing high-level tool integration—especially when using MCP servers—can quickly turn into a configuration nightmare if your underlying agent is too heavy to move between environments.
The Solution: A Truly Personal AI Agent
Nanobot solves this by keeping the core tiny and readable. It provides everything you need for long-running work, including a WebUI, memory management, model routing, and automation, without the overhead of traditional frameworks.
Implementation: Quick Installation
Setting up Nanobot is designed to be seamless. Depending on your operating system, you can use a one-command setup to get up and running immediately.
# For macOS or Linux users:
curl -fsSL https://raw.githubusercontent.com/HKUDS/nanobot/main/scripts/install.sh | sh
# For Windows PowerShell users:
irm https://raw.githubusercontent.com/HKUDS/nanobot/main/scripts/install.ps1 | iexTo ensure a smooth start, you can use the official macOS and Linux installation script or the Windows PowerShell installer.
Initial Setup Steps
- Run the installation command for your OS.
- Initialize your environment by running
nanobot onboard. - Use the
--wizardflag if you prefer an interactive configuration guide.
Configuration: Connecting Your LLMs
Nanobot is model-agnostic. You can connect to cloud providers like Anthropic and OpenAI, or run models locally using tools like LM Studio.
The configuration is handled via a simple config.json file. Use the table below to match your current setup with the correct provider recipe:
| What You Have | Recommended Recipe |
|---|---|
| OpenRouter API Key | OpenRouter Gateway |
| Anthropic or OpenAI Key | Direct Provider Setup |
| Local Ollama Server | Ollama Local Model |
| Custom vLLM/LM Studio | OpenAI-Compatible Provider |
The Result: A Portable Agent Ecosystem
Once configured, you have a fully functional browser-based workbench. You can manage chat sessions, automate tasks, and deploy your agent across different environments without ever losing control of your configuration or data.
Ready to start building? Head over to the Nanobot GitHub repository and clone your new personal AI agent today!