AutoProbeMCP - a browser for your Agent

Allow your Github Copilot Chat to navigate your frontend projects. Also compatible with Chat.fans Agents

javascript

3

GitHub Stars

javascript

Language

4 months ago

First Indexed

3 weeks ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "wladastic-autoprobemcp": {
      "command": "node",
      "args": [
        "/home/yourUserName/mcp-browser-server/build/index.js"
      ]
    }
  }
}

AutoProbeMCP provides a Model Context Protocol server that lets AI assistants automate browser interactions through Playwright. It enables you to navigate pages, fill forms, take screenshots, run scripts, and analyze results in a structured, scriptable way, making web automation and testing with AI agents practical and reproducible.

How to use

Install and run the MCP server, then connect your MCP client (such as Chat.fans agents or GitHub Copilot Chat) to start automating browser tasks. You will launch a browser instance, navigate pages, interact with elements, capture screenshots, run JavaScript in the page context, and retrieve page information and console logs. The server supports multiple browsers and provides a set of tools you can call from your MCP client to perform common automation flows.

How to install

Prerequisites: you need Node.js installed on your system. You will also ensure Playwright browsers are installed for Chromium, Firefox, and WebKit.

# 1) Install dependencies
npm install

# 2) Build the server
npm run build

# 3) Install Playwright browsers (Chromium, Firefox, WebKit)
npx playwright install

# Optional: install system dependencies on Linux
sudo npx playwright install-deps

Additional usage notes

To run locally you will start the MCP server using the command shown in the configuration example. The server is designed to be driven by MCP clients, so you can wire it up in your development environment and begin issuing automation commands such as launching a browser, navigating to URLs, typing into fields, clicking elements, and taking screenshots.

Available tools

launch_browser

Start a new browser instance for automation tasks.

navigate

Navigate the active browser page to a specified URL.

click_element

Click on a page element matching a selector.

type_text

Type text into an input field or form element.

screenshot

Capture a screenshot of the current page or full page.

get_element_text

Extract text content from a page element.

wait_for_element

Wait for a specific element to appear or disappear, with an optional timeout.

evaluate_javascript

Execute custom JavaScript in the browser context and return results.

get_console_logs

Retrieve browser console logs at varying levels (log, info, warn, error, debug).

analyze_screenshot

AI-powered analysis of screenshots using Gemma3 models via Ollama (requires local setup).

get_page_info

Retrieve metadata about the current page, such as URL, title, and viewport.

close_browser

Close the active browser instance to clean up resources.

scroll

Scroll the page in a direction with a specified distance and behavior.

check_scrollability

Check whether vertical or horizontal scrolling is possible and report limits.