Nanobot is an ultra-lightweight personal AI assistant inspired by OpenClaw. It delivers core agent functionality in just ~4,000 lines of code (dynamically verified via bash core_agent_lines.sh) — 99% smaller than Clawdbot’s 430k+ lines. Nanobot supports 12 messaging platforms including WhatsApp, Telegram, Slack, Discord, and WeChat, and is optimized for low-resource deployments on Raspberry Pi and similar hardware.
License: MIT
GitHub: 39.4k+ stars, 6.9k+ forks | HKUDS/nanobot
Origin: Developed by the Data Intelligence Lab at Hong Kong University (HKU) Computer Science department (HKUDS = Hong Kong University Data Science).
- 🪶 Ultra-Lightweight - Just ~4,000 lines of core agent code (99% smaller than Clawdbot, verify with
bash core_agent_lines.sh)
- 🔬 Research-Ready - Clean, readable codebase easy to understand and extend
- ⚡️ Lightning Fast - Minimal footprint, faster startup, lower resource usage
- 💎 Easy-to-Use - One-click deployments with
nanobot onboard
- 📱 12 Messaging Platforms - WhatsApp, Telegram, Slack, Discord, Feishu, DingTalk, QQ, Email, Matrix, Mochat, Wecom, WeChat
- 🔌 MCP Support - Model Context Protocol for tool integration (Stdio and HTTP/SSE transport)
- ⏰ Scheduled Tasks - Cron jobs and heartbeat periodic tasks
- 🧠 Local LLM Support - vLLM integration for local model deployment
- Raspberry Pi Assistant - Run AI agent on low-power hardware
- Multi-Platform Bot - Single bot runtime across WhatsApp, Telegram, Slack, Discord
- Personal AI Agent - 24/7 real-time assistance with scheduled tasks
- Research & Learning - Clean codebase for understanding AI agent architecture
- Team Automation - Multi-channel team communication automation
| Component |
Technology |
| Language |
Python 98.8%, Other 1.2% |
| Python Version |
3.11, 3.12 (requires >=3.11) |
| Package Manager |
pip, uv (recommended) |
| LLM Providers |
20 providers: OpenRouter, Anthropic, OpenAI, Azure OpenAI, DeepSeek, Groq, Gemini, and more |
| Local LLM |
vLLM, Ollama support |
| Messaging |
12 platforms: Telegram, Discord, WhatsApp (Node.js bridge), Slack, Feishu, DingTalk, QQ, Email (IMAP/SMTP), Matrix (E2EE), Mochat, Wecom, WeChat (HTTP long-poll) |
| Container |
Docker, Docker Compose (builds from Dockerfile, no official Docker Hub image) |
| Component |
Minimum |
Recommended |
| OS |
Linux (systemd) |
Ubuntu 20.04+, Debian 10+ |
| Python |
3.11 |
3.12 |
| CPU |
1 core |
2+ cores |
| RAM |
512 MB |
1+ GB |
| Disk |
500 MB |
1+ GB |
| Node.js |
18+ (WhatsApp only) |
20+ (WhatsApp only) |
- ✅ Open-source (MIT License)
- ✅ Active development (v0.1.5 - April 6, 2026)
- ✅ 39.4k+ GitHub stars, 6.9k+ forks
- ✅ PyPI packages:
nanobot-ai (main), nanobot-ai-tng (transitional v0.1.4.post4)
- ✅ Docker support with Dockerfile and docker-compose.yml
- ✅ 12 messaging platform integrations (WeChat added in v0.1.4.post6)
- ✅ MCP (Model Context Protocol) support with Stdio and HTTP/SSE transport
- ✅ Official documentation: nanobot.wiki (available in 6+ languages)
- ⚠️ Security Advisories:
- CVE-2026-2577 (CVSS 10.0 Critical): WhatsApp session hijacking. Versions < v0.1.3.post7 affected. Upgrade immediately.
- GHSA-g6cf-xjwr-pm77 (CVSS 9.0 Critical, no CVE assigned): Authorization bypass via sender ID injection. Versions < v0.1.4.post3 affected. Upgrade to v0.1.4.post4 or higher.
- CVE-2026-33654 (High, GHSA-4gmr-2vc8-7qh3): Zero-click indirect prompt injection and authentication bypass via email polling. Published March 27, 2026. Researcher: Re-bin. Upgrade to latest version.
Any questions?
Feel free to contact us. Find all contact information on our contact page.