DrissionPage

You can run DrissionPage MCP Server to automate browser tasks, scrape data, and convert web pages to Markdown for AI models. It provides a single browser instance you control through the MCP interface, making it easy to perform navigation, form filling, data extraction, and content transformation from your AI workflows.

How to use

Use an MCP client to call the available tools exposed by the server. You can start a browser, navigate to pages, interact with elements, extract data, convert pages to Markdown, and save results. The server is designed for sequential tasks using a single browser instance and includes debugging visibility by default in headful mode.

How to install

Prerequisites: Ensure Python is installed on your system and you have network access to install dependencies.

# 1) Install Python dependencies for the MCP server
cd drissionpage_mcp
pip install -r requirements.txt

# 2) Run the MCP server (stdio mode) using the documented command
python server.py

Configuration and usage notes

Configure your MCP client to connect to the stdio server using the exact command path shown. The example demonstrates how to invoke the server from a client configuration.

{
  "mcpServers": {
    "drissionpage": {
      "command": "python",
      "args": ["/Users/Desktop/drissionpage_mcp/server.py"]
    }
  }
}

Usage examples

Example 1: Visit a page and save as Markdown

# 1. init_browser()
# 2. navigate(url="https://news.ycombinator.com")
# 3. page_to_markdown(file_path="hn_news.md")
# 4. close_browser()

扩展用法与常见场景

Common tasks you can perform include filling forms, extracting tables, handling infinite scroll, and managing cookies. You can also switch tabs, take screenshots, and execute custom JavaScript. The core features focus on page-to-Markdown conversion, structured data extraction, and browser automation through a unified toolset.

Available tools

init_browser

Initialize a single browser instance for automated tasks.

get_browser_status

Query the current status of the browser instance.

close_browser

Close the active browser and release resources.

navigate

Navigate the browser to a specified URL.

find_elements

Find elements on the page using selectors.

click_element

Click a specified page element.

input_text

Type text into a target input field.

get_element_text

Get the text content of a page element.

get_element_attribute

Retrieve an attribute value from a page element.

wait_for_element

Wait for an element to appear with a timeout.

scroll_page

Scroll the page programmatically.

take_screenshot

Capture a screenshot of the current page.

execute_javascript

Run custom JavaScript in the page context.

page_to_markdown

Convert the current page content to Markdown and save to a file.

extract_table_data

Extract data from tables and return structured formats.

smart_extract

Perform structured data extraction across multiple elements.

fill_form

Automatically fill form fields and submit.

handle_infinite_scroll

Handle pages that load content as you scroll.

manage_cookies

Manage cookies for sessions and authentication.

switch_to_tab

Switch the browser focus between tabs.