mcp -puppeteer-py

This MCP server provides browser automation capabilities via Playwright in Python. It exposes tools that let an LLM-driven agent open web pages, navigate, take full-page or element-specific screenshots, interact with forms and UI elements, and execute JavaScript within a real browser. Key tools include puppeteer_navigate for loading URLs, puppeteer_screenshot for capturing visuals, puppeteer_click and puppeteer_fill for interacting with page elements, and puppeteer_evaluate for running scripts in the page context. The server outputs structured results and captures console logs to help diagnose issues during browser automation tasks.

Prerequisites:

• Python 3.8+
• pip (Python package installer)

Installation steps:

1. Create and activate a Python virtual environment (optional but recommended): python -m venv venv source venv/bin/activate # On Windows use: venv\Scripts\activate

2. Install required packages: pip install -r requirements.txt

3. Install Playwright browsers: playwright install

4. Run the server: python puppeteer_server.py

Notes:

• Ensure that the path the MCP uses to puppeteer_server.py matches your project structure.
• If you modify dependencies, update requirements.txt accordingly.

Tips and considerations:

• Browser launches may be non-headless by default for debugging; verify your configuration if headless behavior is required.
• The default viewport is 1280x720; adjust as needed for your tasks.
• Timeouts are configurable per tool invocation—tune them to balance responsiveness and reliability in your environment.
• Console logs from the automated browser are captured for analysis; inspect them when failures occur.
• Screenshots are stored in memory and can be returned as base64-encoded data; adjust storage strategy if you need persistent artifacts.
• If you encounter navigation or element-not-found errors, double-check selectors and page readiness before retrying.