HKUDS/nanobot
Ultra-lightweight, open-source, self-hosted personal AI agent framework in Python with WebUI, tools, memory, MCP, multi-agent workflows, automation, and chat apps
46,702 ★8,263 forksPythonUpdated 22h ago
What you need to know
Ultra-lightweight, open-source, self-hosted personal AI agent framework written in Python, running in a WebUI, terminal, or chat apps. Combines tools, long-term memory, MCP integrations, model routing, multi-agent delegation, scheduled automation, and an OpenAI-compatible API in a small, readable core.
Install
curl -fsSL https://raw.githubusercontent.com/HKUDS/nanobot/main/scripts/install.sh | sh (macOS/Linux) irm https://raw.githubusercontent.com/HKUDS/nanobot/main/scripts/install.ps1 | iex (Windows PowerShell) uv tool install nanobot-ai python -m pip install nanobot-ai From source: git clone https://github.com/HKUDS/nanobot.git && cd nanobot && python -m pip install . (requires bun or npm)
Usage
- •nanobot webui — recommended first run; opens http://127.0.0.1:8765 and lets you configure provider/model in Settings → Models
- •nanobot webui --background / nanobot gateway — keep the full gateway (channels, automations) running after the terminal closes
- •nanobot agent / nanobot agent -m "Hello!" — interactive terminal chat or one-shot request
- •Connect chat apps (Telegram, Discord, Slack, WeChat, Feishu, Email, Mattermost, Teams) for chat-native reach
Key features
- ✓Runs in a browser WebUI or terminal and connects to chat apps (Telegram, Discord, Slack, WeChat, Email, Mattermost, Teams, Feishu)
- ✓Built-in tools: files, shell, web search, web fetch, MCP, cron, image generation, and subagents
- ✓Long-term memory via 'Dream', plus long-horizon goals and scheduled automations
- ✓Model freedom: OpenAI-compatible APIs, local LLMs (Ollama, vLLM), image generation, search, and fallback models
- ✓Exposes a Python SDK and OpenAI-compatible API; one-click deploy to Render, or Docker / Linux service / macOS LaunchAgent self-hosting
Best for
A self-hosted personal AI assistant needing persistent long-running workflows, chat-app reach, and full model freedom in one lightweight Python runtime.
Caveats
- ⚠Requires Python 3.11 or newer
- ⚠Source install requires bun or npm to build the WebUI; pip may report externally-managed-environment on macOS/Linux (use uv/pipx/venv instead)
- ⚠First-run WebUI binds to localhost by default and is not exposed to your LAN
- ⚠PyPI install is the stable track; newest features/experiments require installing from source (--dev)
- ⚠Render deployment needs a paid Render service for persistent disks (sessions, memory, WebUI history)
Platforms: macOS · Linux · WindowsClients: WebUI · Terminal · Telegram · Discord · Slack · WeChat · Feishu · Email · Mattermost · Teams
Documentation ↗Reviewed 2026-08-07
Topics
agent-frameworkai-agentai-agentschatbotchatopsdiscord-botllm-agentsllmslocal-llmmcpmodel-context-protocolmulti-agentopenai-compatibleopenclawpersonal-ai-assistantpythonself-hostedtelegram-bot-ai-assistantwebuiworkflow-automation
- Stars
- 46,702★
- Forks
- 8,263
- Language
- Python
- License
- MIT
- Created
- 2026-02-01
- Last push
- 2026-08-06