🐈 nanobot
The ultra-lightweight personal AI assistant framework inspired by OpenClaw
Why nanobot?
nanobot delivers core agent functionality with 99% fewer lines of code than OpenClaw. It’s designed to be:Ultra-Lightweight
Just ~4,000 lines of core agent code — minimal footprint means faster startup and quicker iterations
Research-Ready
Clean, readable code that’s easy to understand, modify, and extend for research projects
Lightning Fast
Lower resource usage and minimal overhead for production deployments
Easy to Use
One-click deployment — initialize, configure, and start chatting in under 2 minutes
Key Features
Multi-Provider LLM
Support for 15+ LLM providers including OpenRouter, Anthropic, OpenAI, DeepSeek, and more
10+ Chat Platforms
Telegram, Discord, WhatsApp, Feishu, Slack, DingTalk, QQ, Email, Matrix, and Mochat
Extensible Skills
Compatible with OpenClaw skills — GitHub, weather, web search, summarization, and more
MCP Support
Model Context Protocol integration for external tool servers
Built-in Tools
Shell execution, filesystem operations, web fetching, message routing, and subagent spawning
Persistent Memory
Session management and long-term memory storage
Scheduled Tasks
Cron-based scheduling and periodic heartbeat tasks
OAuth Providers
Support for OpenAI Codex and GitHub Copilot with OAuth authentication
Secure by Default
Workspace sandboxing and security controls built-in
Quick Example
Get up and running in three simple steps:What’s Next?
Installation Guide
Detailed installation instructions for all platforms
Quickstart Tutorial
Get your first agent running in 2 minutes
Architecture Overview
Learn how nanobot works under the hood
Configuration Guide
Configure providers, channels, and security settings
Chat Channels
Connect nanobot to Telegram, Discord, WhatsApp, and more
Skills System
Extend nanobot with custom skills and tools
Community
GitHub
View source code and contribute
Discord
Join our Discord community
PyPI
View package on PyPI