Tribute to nanobot
This project is built upon nanobot for secondary development. nanobot was inspired by Clawdbot, implementing complete Agent functionality with approximately 4000 lines of core code. Thanks to the nanobot community for the lightweight, readable, and extensible foundation.

This fork extends nanobot with Web UI, MCP support, additional channels (Discord, QQ, DingTalk), document/office skills, system monitoring, and more.

nanobot-webui comes with a powerful skill system that extends the AI's capabilities:

The skill system is extensible. You can create custom skills by implementing skill functions that the AI can invoke. See the skill-creator skill for a template.

nanobot-webui supports scheduled tasks through a built-in Cron system, allowing you to set up recurring reminders and automated actions:

# Set a daily reminder
"每天早上 9 点提醒我喝水"
# Set a weekly meeting reminder
"每周一上午 10 点提醒我开周会"
# Set an interval check
"每半小时检查一下服务器状态"

The Cron system runs in the background and automatically delivers scheduled messages to your configured channels.

Multi-turn conversation with Nanobot, Markdown rendering, new sessions and history. Example shows weather query — Agent fetches real-time data via MCP tools and returns structured card results.

Chat interface

View service health, uptime, active session count, installed Skills, workspace path, and system environment info.

System status

Manage IM channel enablement and config — Feishu, Discord, QQ, DingTalk, Telegram, WhatsApp, and more.

Channels config

Manage AI model providers — DeepSeek, Qwen, Zhipu, OpenAI, Anthropic, OpenRouter, vLLM, Groq, Gemini, etc. Add, edit, or remove providers.

Providers config

Set the default Agent model — model name (provider/model-name format), Temperature, Max Tokens, and other parameters.

Default model config

Manage Model Context Protocol servers — stdio, http, sse, streamable_http. Import/generate JSON or add new MCP services.

MCP config

Manage AI skills — select skill folders for upload to workspace, view installed skills (version, status, description), enable or disable.

Skills management

git clone https://github.com/codemo1991/nanobot-webui.git
cd nanobot-webui
pip install -e .

The intelligent startup scripts in the scripts/ directory handle everything automatically — the simplest way to get started:

Supported Platforms:

• Linux / macOS: nanobot-launcher.sh
• Windows PowerShell: nanobot-launcher.ps1

What it does:

• ✅ Automatically checks Python (≥3.11) and Node.js environment
• ✅ Automatically installs/updates Python dependencies
• ✅ Automatically builds the frontend
• ✅ One-click launches the Web UI service

Usage:

• Windows (PowerShell):

  .\scripts\nanobot-launcher.ps1

• Linux / macOS:

  chmod +x scripts/nanobot-launcher.sh
  ./scripts/nanobot-launcher.sh

Then open http://127.0.0.1:6788

Tip: Run the script again anytime to update dependencies and restart the service — the simplest way to keep your installation up-to-date!

# Build frontend (first time)
cd web-ui && npm install && npm run build && cd ..
# Start Web UI (backend + static files)
nanobot web-ui

Open http://127.0.0.1:6788 — Chat, Config, and System pages are available.

nanobot onboard # Initialize config
# Edit ~/.nanobot/config.json with your API key
nanobot agent -m "Hello!"

# Build the image
docker build -t nanobot-webui .
# Run (recommended: mount volume for config persistence)
docker run -d -p 6788:6788 -v nanobot-data:/root/.nanobot --name nanobot nanobot-webui
# Or use host path for config
docker run -d -p 6788:6788 -v ~/.nanobot:/root/.nanobot --name nanobot nanobot-webui

Then open http://127.0.0.1:6788. On first launch, the app auto-creates ~/.nanobot/config.json; add your API key via the Config page in the Web UI.

• Chat — Create sessions, multi-turn conversations, Markdown rendering, session persistence
• File Browser — Browse and manage workspace files directly from the UI
• Session History — View and manage conversation history
• Model Switching — Easily switch between different AI models
• Config — Manage Channels (IM), Providers, Models, MCP servers, Skills
• System — Health check, uptime, session count, system info, config export

# Feishu
pip install nanobot-ai[feishu]
# QQ
pip install nanobot-ai[qq]
# DingTalk
pip install nanobot-ai[dingtalk]
# MCP
pip install nanobot-ai[mcp]

nanobot/
├── agent/ # Core agent (loop, context, memory, tools)
│ └── tools/ # mcp.py, filesystem, shell, registry...
├── channels/ # telegram, whatsapp, feishu, discord, qq, dingtalk
├── web/ # REST API server (api.py)
├── mcp/ # MCP loader
├── storage/ # StatusRepository (SQLite)
├── services/ # SystemStatusService
├── skills/ # claude-code, git-manager, xlsx, pdf, pptx, skill-creator, mirror-system, code-review-expert...
├── providers/ # LLM providers (OpenAI, Anthropic, DeepSeek, etc.)
├── config/ # Extended schema (Discord, QQ, DingTalk, MCP)
└── cli/ # web-ui command, status, etc.
web-ui/ # React SPA (Chat, Config, System)
scripts/ # Startup scripts (nanobot-launcher.sh, nanobot-launcher.ps1)

Contributions are welcome! Please feel free to submit a Pull Request or open an Issue for bugs and feature requests.

# Clone the repository
git clone https://github.com/codemo1991/nanobot-webui.git
cd nanobot-webui
# Install in development mode
pip install -e .
# Install web-ui dependencies
cd web-ui && npm install
# Run development server
cd ..
nanobot web-ui

• Python: Follow PEP 8
• JavaScript/TypeScript: Follow ESLint configuration
• Commit messages: Use clear, descriptive messages

MIT License — see LICENSE file for details.

• nanobot — The base project this fork builds upon
• nanobot contributors — For the lightweight, research-friendly design

Thanks for visiting ✨ nanobot-webui!