nanobot-webui 0.1.3

Web management panel for nanobot

Nanobot WebUI

English | 中文

---

A self-hosted web management panel for nanobot (PyPI) — a multi-channel AI agent framework.
Provides a full-featured UI to configure, converse with, and manage your nanobot instance, with no modifications to the core library.

---

Table of Contents

• Features
• Quick Start
  • pip install (recommended)
  • Docker
• CLI Reference
• Development
• Architecture
• Authentication
• Tech Stack

---

Screenshots

Login	Dashboard	Chat

---

Features

Module	Description
Dashboard	Channel health, session / skill / cron statistics at a glance
Chat	Real-time conversation with the agent over WebSocket
Providers	Configure API keys & base URLs for OpenAI, Anthropic, DeepSeek, Azure, and more
Channels	View and configure all IM channels (Telegram, Discord, Feishu, DingTalk, Slack, QQ, WhatsApp, Email, Matrix, MoChat)
MCP Servers	Manage Model Context Protocol tool servers
Skills	Enable / disable agent skills; edit workspace skills in-browser
Cron Jobs	Schedule, edit, and toggle recurring tasks
Agent Settings	Model, temperature, max tokens, memory window, workspace path, etc.
Users	Multi-user management with admin / user roles

---

Quick Start

pip install (recommended)

pip install nanobot-webui

The pre-built React frontend is bundled in the wheel — no Node.js required.
After installation, use the nanobot command to start the WebUI:

# Foreground (WebUI + gateway combined)
nanobot webui

# Custom port
nanobot webui --port 9090

# Background daemon (recommended for long-running deployments)
nanobot webui --daemon

Open http://localhost:8080 — default credentials: admin / nanobot — change on first login.

---

Docker

Prerequisites: Docker ≥ 24 with the Compose plugin (docker compose).

Option 1 — Docker Compose (recommended)

Create a docker-compose.yml:

services:
  webui:
    image: kangkang223/nanobot-webui:latest
    container_name: nanobot-webui
    volumes:
      - ~/.nanobot-webui:/root/.nanobot   # config & data persistence
    ports:
      - "8080:8080"    # WebUI
      - "18790:18790"  # nanobot gateway (optional, for IM channel webhooks)
    restart: unless-stopped

Then:

# Pull the latest image and start in background
docker compose up -d

# View logs
docker compose logs -f

# Stop
docker compose down

Open http://localhost:8080 — default credentials: admin / nanobot.

Data directory: all config, sessions, and workspace files are stored in ~/.nanobot-webui on the host (mapped to /root/.nanobot inside the container).

Option 2 — Build from source

git clone https://github.com/Good0007/nanobot-webui.git
cd nanobot-webui

# Build the multi-stage image (bun build → python runtime)
docker build -t nanobot-webui .

# Run
docker run -d \
  --name nanobot-webui \
  -p 8080:8080 \
  -v ~/.nanobot-webui:/root/.nanobot \
  --restart unless-stopped \
  nanobot-webui

Option 3 — Makefile shortcuts

If you have the repository cloned, the bundled Makefile wraps common tasks:

make up           # docker compose up -d
make down         # docker compose down
make logs         # follow compose logs
make restart      # docker compose restart
make build        # build local single-arch image
make release-dated  # build & push :YYYY-MM-DD + :latest (multi-arch)

---

CLI Reference

Installing nanobot-webui extends the nanobot command with the following subcommands:

nanobot webui — Start the WebUI

Usage: nanobot webui [OPTIONS] [COMMAND]

Options:
  -p, --port INTEGER        WebUI HTTP port  (default: 8080)
  -g, --gateway-port INT    nanobot gateway port  (default: from config)
      --host TEXT           Bind address  (default: 0.0.0.0)
  -w, --workspace PATH      Override workspace directory
  -c, --config PATH         Path to config file
      --no-gateway          Start WebUI only; skip nanobot gateway/agent
  -d, --daemon              Run in background; return immediately

nanobot webui                          # foreground (WebUI + gateway)
nanobot webui --port 9090              # custom port
nanobot webui --daemon                 # background daemon
nanobot webui --daemon --port 9090     # background + custom port
nanobot webui --no-gateway             # WebUI only (gateway running elsewhere)
nanobot webui --workspace ~/myproject  # override workspace

nanobot webui logs — View logs

Usage: nanobot webui logs [OPTIONS]

Options:
  -f, --follow          Stream log output in real time (like tail -f)
  -n, --lines INTEGER   Number of lines to show  (default: 50)

nanobot webui logs              # last 50 lines
nanobot webui logs -f           # stream in real time
nanobot webui logs -f -n 100    # stream, show last 100 lines

Log file: ~/.nanobot/webui.log

nanobot stop — Stop the background service

nanobot stop    # sends SIGTERM; force-kills after 6 s if needed

nanobot status — Show runtime status

nanobot status  # shows WebUI process info + nanobot config summary

Example output:

🐈 nanobot Status

WebUI: ✓ running (PID 12345 • http://localhost:8080)
Log  : /home/user/.nanobot/webui.log

Config: /home/user/.nanobot/config.json ✓
Workspace: /home/user/.nanobot/workspace ✓
Model: gpt-4o
...

State files: PID → ~/.nanobot/webui.pid, port → ~/.nanobot/webui.port

---

Development

Prerequisites: Python ≥ 3.11, Bun ≥ 1.0

# 1. Clone and install backend in editable mode
git clone https://github.com/Good0007/nanobot-webui.git
cd nanobot-webui
pip install -e .

# 2. Start the backend
nanobot webui                        # API + static on :8080

# 3. Start the frontend dev server (separate terminal)
cd web
bun install
bun dev                              # http://localhost:5173  (proxies /api → :8080)

To produce a production build:

cd web
bun run build          # outputs to web/dist/, setup.py copies it to webui/web/dist/
cd ..
nanobot webui          # backend now serves webui/web/dist/ as static files

---

Architecture

nanobot-webui/
├── webui/                      # Python package (importable as `webui`)
│   ├── __init__.py
│   ├── __main__.py             # Entry point: python -m webui
│   ├── web/
│   │   └── dist/               # Built React assets (generated by bun run build)
│   └── api/                    # FastAPI backend
│       ├── auth.py             # JWT + bcrypt helpers
│       ├── users.py            # UserStore  (~/.nanobot/webui_users.json)
│       ├── deps.py             # FastAPI dependency injection
│       ├── gateway.py          # ServiceContainer + server lifecycle
│       ├── server.py           # FastAPI app factory (static serving, SPA fallback)
│       ├── channel_ext.py      # ExtendedChannelManager (non-invasive subclass)
│       ├── models.py           # Pydantic response models
│       └── routes/             # One file per domain
│           ├── auth.py         #   POST /api/auth/login|register|change-password
│           ├── channels.py     #   GET|PATCH /api/channels
│           ├── config.py       #   GET|PATCH /api/config
│           ├── cron.py         #   CRUD /api/cron
│           ├── mcp.py          #   GET|PATCH /api/mcp
│           ├── providers.py    #   GET|PATCH /api/providers
│           ├── sessions.py     #   GET|DELETE /api/sessions
│           ├── skills.py       #   GET|POST /api/skills
│           ├── users.py        #   CRUD /api/users  (admin only)
│           └── ws.py           #   WebSocket /ws/chat
├── web/                        # React 18 + TypeScript frontend source
│   ├── src/
│   │   ├── pages/              # One component per route
│   │   ├── components/         # Layout, chat, shared UI
│   │   ├── hooks/              # TanStack Query data hooks
│   │   ├── stores/             # Zustand stores (auth, chat)
│   │   ├── lib/                # axios instance, WS manager, utils
│   │   └── i18n/               # zh / en JSON translation files
│   ├── eslint.config.js
│   └── package.json
├── Dockerfile                  # Multi-stage: bun build → python runtime
├── docker-compose.yml
├── pyproject.toml
└── setup.py                    # Build hook: runs bun run build, copies dist into webui/

Design principle: the backend is entirely non-invasive — it imports nanobot libraries but never patches their source. Runtime monkey-patches (applied in __main__.py) are limited to quality-of-life tweaks (e.g. treating an empty allow_from list as "allow all").

---

Authentication

Detail	Value
Default credentials	admin / nanobot
Credential storage	~/.nanobot/webui_users.json (bcrypt-hashed passwords)
Token type	JWT (HS256)
Token expiry	7 days
JWT secret	Auto-generated per instance, stored in ~/.nanobot/webui_secret.key

Security note: change the default password immediately after first login.

---

Tech Stack

Layer	Library / Tool
Backend framework	FastAPI + Uvicorn
Auth	PyJWT + bcrypt
Frontend framework	React 18 + TypeScript + Vite
UI components	shadcn/ui + Tailwind CSS v3
Client state	Zustand (persist middleware)
Server state	TanStack Query v5
i18n	react-i18next (zh / en)
Theme	next-themes (light / dark / system)
Real-time	WebSocket (/ws/chat)
Package manager	Bun