LAM-mcp

Gafoor2005/LAM-mcp

3.2

If you are the rightful owner of LAM-mcp and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

The Browser Automation MCP Server provides intelligent browser automation with session-based context analysis, enabling AI assistants to interact with web pages and make smart decisions.

Tools
5
Resources
0
Prompts
0

Browser Automation MCP Server with Session Context

A Model Context Protocol (MCP) server that provides intelligent browser automation with session-based context analysis. This server enables AI assistants to interact with web pages and make smart decisions based on page analysis and surrounding context.

✨ Key Features

🧠 Session-Based Context Intelligence

  • Smart Element Identification: Analyzes page structure to find the right elements for your task
  • Context-Aware Suggestions: Recommends elements based on labels, nearby text, and relevance
  • Semantic Search: Find relevant page sections using natural language queries
  • Progress Tracking: Monitor actions and success rates during the session
  • Ephemeral Memory: Session-only storage that clears when done (no persistent learning)
  • Relevance Scoring: Elements ranked by confidence based on context analysis

🌐 Core Browser Automation

  • Web Navigation: Navigate to URLs, handle redirects, and manage browser history
  • Element Interaction: Click buttons, links, and other interactive elements
  • Text Input: Type into form fields, search boxes, and text areas
  • Content Extraction: Extract text, HTML, and structured data from web pages
  • Screenshot Capture: Take full-page or element-specific screenshots
  • JavaScript Execution: Run custom JavaScript code in the browser context

🚀 Advanced Capabilities

  • Form Handling: Automatically fill out complex web forms
  • Cookie Management: Get, set, and manage browser cookies
  • Session Persistence: Maintain browser sessions across operations
  • Multi-Browser Support: Chrome, Firefox, Edge with automatic driver management
  • Headless/Headed Modes: Run browsers with or without GUI
  • Element Waiting: Smart waiting for dynamic content and AJAX loads

📦 Installation

  1. Install Dependencies:

    pip install -e .
# or
uv add mcp selenium webdriver-manager pillow beautifulsoup4 chromadb sentence-transformers

  2. Browser Drivers (handled automatically by webdriver-manager):

    • Chrome/Chromium driver
    • Firefox GeckoDriver
    • Edge WebDriver

  3. Context Analysis Dependencies:

    • ChromaDB for in-memory vector storage
    • sentence-transformers for semantic embeddings (all-MiniLM-L6-v2)

Usage

Running the Server

# STDIO transport (for local development)
python -m browser_mcp_server.server

# SSE transport (for web-based clients)
python -m browser_mcp_server.server --transport sse --port 8000

Configuration with Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "browser-automation": {
      "command": "uv",
      "args": ["run", "python", "-m", "browser_mcp_server.server"],
      "env": {
        "BROWSER_HEADLESS": "true",
        "BROWSER_TYPE": "chrome"
      }
    }
  }
}

Available Tools

🧠 Session Context Intelligence Tools

  • analyze_current_page: Analyze page structure and store in session context
  • find_page_context: Find relevant page sections for your task using semantic search
  • get_smart_element_selector: Get elements with surrounding context for better identification
  • track_action_result: Track actions and success/failure in current session
  • get_session_progress: View session statistics, actions taken, and success rate
  • clear_session_context: Clear session memory and start fresh

Navigation & Page Control

  • navigate_to_url: Navigate to a specific URL
  • get_page_source: Get the full HTML source of the current page
  • get_rendered_html: Get dynamically rendered HTML (after JavaScript)
  • get_current_url: Get the current page URL
  • refresh_page: Refresh the current page
  • go_back: Navigate back in browser history
  • go_forward: Navigate forward in browser history

Element Interaction

  • click_element: Click on elements by CSS selector or XPath
  • type_text: Type text into input fields
  • clear_field: Clear text from input fields
  • get_element_text: Extract text content from elements
  • get_element_attribute: Get attribute values from elements
  • is_element_present: Check if an element exists on the page

Content Extraction

  • extract_links: Extract all links from the current page
  • extract_images: Extract all images and their attributes
  • extract_tables: Extract table data as structured JSON
  • extract_forms: Extract form fields and their properties
  • search_text: Search for text patterns on the page

Form Automation

  • fill_form: Fill out entire forms with structured data
  • select_dropdown_option: Select options from dropdown menus
  • check_checkbox: Check or uncheck checkbox elements
  • select_radio_button: Select radio button options

Visual & Media

  • take_screenshot: Capture full-page or element screenshots
  • scroll_page: Scroll the page in various directions
  • hover_element: Hover mouse over elements
  • drag_and_drop: Drag and drop elements

JavaScript & Advanced

  • execute_javascript: Execute custom JavaScript code
  • wait_for_element: Wait for elements to appear or disappear
  • wait_for_page_load: Wait for page load completion
  • switch_to_frame: Switch context to iframes
  • handle_alert: Handle JavaScript alerts and confirmations

Session & Cookies

  • get_cookies: Retrieve browser cookies
  • set_cookies: Set browser cookies
  • delete_cookies: Delete specific or all cookies
  • get_local_storage: Access browser local storage
  • set_local_storage: Set local storage values

Environment Variables

  • BROWSER_TYPE: Browser to use (chrome, firefox, edge) - default: chrome
  • BROWSER_HEADLESS: Run in headless mode (true/false) - default: true
  • BROWSER_TIMEOUT: Default timeout for operations in seconds - default: 30
  • BROWSER_WINDOW_SIZE: Browser window size (1920x1080) - default: 1920x1080
  • BROWSER_USER_AGENT: Custom user agent string
  • BROWSER_DOWNLOAD_DIR: Directory for file downloads

Example Usage Patterns

Web Scraping

# Navigate to a website and extract data
await navigate_to_url("https://example.com")
await wait_for_element("css:.content")
content = await extract_text("css:.main-content")
links = await extract_links()

Form Automation

# Fill out and submit a contact form
await navigate_to_url("https://example.com/contact")
await fill_form({
    "name": "John Doe",
    "email": "john@example.com",
    "message": "Hello from MCP!"
})
await click_element("css:button[type='submit']")

E-commerce Automation

# Search and interact with e-commerce sites
await navigate_to_url("https://shop.example.com")
await type_text("css:#search-input", "laptop")
await click_element("css:.search-button")
await wait_for_element("css:.product-list")
products = await extract_text("css:.product-item")

Security Considerations

This server provides powerful browser automation capabilities that can:

  • Navigate to any website
  • Execute arbitrary JavaScript
  • Access and modify cookies and local storage
  • Take screenshots of web content
  • Fill out and submit forms

Recommendations:

  • Use in controlled environments
  • Implement proper authentication for production
  • Review and approve automation scripts
  • Monitor for suspicious activities
  • Use headless mode in production
  • Implement rate limiting and timeouts

Development

Running Tests

uv run pytest tests/

Code Formatting

uv run black browser_mcp_server/
uv run isort browser_mcp_server/

Type Checking

uv run mypy browser_mcp_server/

License

MIT License - see LICENSE file for details.