AI自动化代理

graph TD
    A[MCP Agent] --> B[LangChain]
    A --> C[MCP Protocol]
    A --> D[Playwright]
    A --> E[Groq LLM]
    
    B --> F[Python 3.13+]
    C --> G[MCP Servers]
    D --> H[Browser Control]
    E --> I[AI Models]
    
    style A fill:#ff6b6b,stroke:#333,stroke-width:4px
    style B fill:#4ecdc4,stroke:#333,stroke-width:2px
    style C fill:#45b7d1,stroke:#333,stroke-width:2px
    style D fill:#96ceb4,stroke:#333,stroke-width:2px
    style E fill:#ffeaa7,stroke:#333,stroke-width:2px

# Browse and search
you: open google.com and search for "weather in Tokyo"

# Take screenshots
you: navigate to github.com and take a screenshot

# Fill forms
you: go to example.com/contact and fill out the form with name "John" and email "john@example.com"

# Extract data
you: visit amazon.com and find the price of the latest iPhone

# Multi-step workflows
you: book a flight from New York to London for next week, find the cheapest option

# Data collection
you: visit reddit.com/r/technology and get the top 5 post titles

# Testing
you: test the login functionality of example.com with username "test" and password "test123"

# Required
GROQ_API_KEY=your_groq_api_key_here

# Optional - Additional LLMs
OPENAI_API_KEY=your_openai_key_here
GOOGLE_API_KEY=your_google_key_here

# Optional - MCP Server Tokens
GITHUB_PERSONAL_ACCESS_TOKEN=your_github_token

config = {
    "mcpServers": {
        "playwright": {
            "command": "npx",
            "args": ["-y", "@playwright/mcp@latest"]
        },
        "filesystem": {
            "command": "npx",
            "args": ["-y", "@modelcontextprotocol/server-filesystem", "/tmp"]
        },
        "github": {
            "command": "npx",
            "args": ["-y", "@modelcontextprotocol/server-github"],
            "env": {"GITHUB_PERSONAL_ACCESS_TOKEN": "your_token"}
        }
    }
}

# Run from project directory
uv run app.py

# Install globally
pip install -e .
mcp-agent

FROM python:3.13-slim
WORKDIR /app
COPY --from=ghcr.io/astral-sh/uv:latest /uv /bin/uv
COPY . .
RUN uv sync --frozen
CMD ["uv", "run", "app.py"]

docker build -t mcp-agent .
docker run -it --env-file .env mcp-agent

mcpwheather/
├── 📄 app.py              # Main application entry point
├── 📄 pyproject.toml      # Project configuration
├── 📄 .env                # Environment variables
├── 📄 .python-version     # Python version
├── 📄 uv.lock             # Dependency lock file
├── 📄 README.md           # Project documentation
└── 📄 .gitignore          # Git ignore rules

# Check dependencies
uv tree

# Test API connection
python -c "from langchain_groq import ChatGroq; print('✅ API OK')"

# Verify MCP server
npx @playwright/mcp@latest --help

# Clone your fork
git clone https://github.com/yourusername/mcpwheather.git
cd mcpwheather

# Set up development environment
python -m venv .venv
source .venv/bin/activate
uv sync --dev

# Run tests
uv run pytest

# Start coding! 🚀