Documentation Center

Requirements

• Python 3.10 or higher
• At least 500MB of available disk space
• At least one LLM API Key (DeepSeek, OpenAI, etc.)

Install WeClaw

Install with a single pip command, no complex configuration needed:

pip install weclawpy

First Run

After installation, choose your preferred way to start WeClaw:

# CLI 模式
python -m weclaw

# GUI 模式
weclaw-gui

Chat Features

WeClaw supports natural language conversations with multiple LLMs, featuring:

Tool Invocation

Describe your task in natural language, and WeClaw will automatically select and invoke the appropriate tools:

# 使用自然语言调用工具
> 帮我截取当前屏幕
> 搜索最近修改的文件
> 打开浏览器访问 github.com

Workflow Configuration

With the workflow feature, you can define multi-step automation tasks for AI to execute according to preset processes.

Model Configuration

WeClaw supports configuring multiple LLM models simultaneously in config/models.toml:

# config/models.toml
[deepseek]
api_key = "sk-xxx"
model = "deepseek-chat"

[openai]
api_key = "sk-xxx"
model = "gpt-4o"
base_url = "https://api.openai.com/v1"

[claude]
api_key = "sk-ant-xxx"
model = "claude-3-5-sonnet-20241022"

Tool Customization

You can enable or disable specific tools and configure their parameters:

# config/tools.json
{
  "enabled_tools": ["shell", "file", "browser", "voice"],
  "tool_settings": {
    "browser": {
      "headless": true,
      "timeout": 30000
    }
  }
}

System Prompt Customization

Customize AI behavior and personality through the system prompt settings to match your usage habits.

Remote Service Overview

WeClaw remote service provides WebSocket and REST interfaces for remote desktop AI control from PWA mobile or other clients.

WebSocket Communication

WebSocket interface for real-time bidirectional communication, supporting streaming conversations and live status updates:

// WebSocket 连接
const ws = new WebSocket('ws://localhost:8765/ws');

ws.onopen = () => {
  ws.send(JSON.stringify({
    type: 'chat',
    message: 'Hello, WeClaw!'
  }));
};

ws.onmessage = (event) => {
  const data = JSON.parse(event.data);
  console.log('Response:', data);
};

REST Endpoints

REST API provides the following main endpoints:

Local Deployment

The simplest deployment method, suitable for personal use:

# 本地部署
pip install weclawpy
weclaw-gui

Remote Service Deployment

If you need remote control from your phone, start the remote service:

# 启动远程服务
python -m weclaw.remote_server --host 0.0.0.0 --port 8765

# 配置防火墙（可选）
ufw allow 8765/tcp

Docker Deployment

Use Docker to quickly deploy a complete WeClaw environment:

# docker-compose.yml
version: '3.8'
services:
  weclaw:
    image: weclaw/weclaw:latest
    ports:
      - "8765:8765"
    volumes:
      - ./config:/app/config
      - ./data:/app/data
    environment:
      - DEEPSEEK_API_KEY=sk-xxx