Concurrent Browser Mcp

Name: Concurrent Browser Mcp
Author: sailaoda

@sailaoda

Repository

About Concurrent Browser Mcp

A concurrent browser MCP server that supports multiple parallel browser instances.

#browser #concurrent #parallel

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "concurrent-browser-mcp": {
      "command": "npx",
      "args": [
        "concurrent-browser-mcp",
        "--max-instances",
        "20"
      ]
    }
  }
}

Tools

Create a new browser instance

List all browser instances

Close the specified browser instance

Close all browser instances

Navigate to a specified URL

Go back to the previous page

Go forward to the next page

Refresh the current page

Click on a page element

Type text into an element

Fill a form field

Select an option from a dropdown

Get detailed page information including full HTML content, page statistics, and metadata

Get element text content

Get element attribute value

Take a screenshot of the page or element

Wait for an element to appear

Wait for page navigation to complete

Execute JavaScript code in the page context

Get page content in Markdown format, optimized for large language models

Overview

What is Concurrent Browser MCP?

Concurrent Browser MCP is an MCP (Model Context Protocol) server built on Playwright that supports managing multiple browser instances concurrently. It is designed for developers and AI agents that need scalable, isolated, and automated browser interactions (navigation, clicking, form filling, screenshots, etc.) across Chromium, Firefox, and WebKit.

How to use Concurrent Browser MCP?

Install via npm install -g concurrent-browser-mcp or run directly with npx concurrent-browser-mcp. Configure in your MCP client using the provided JSON examples under mcpServers. Launch with command-line options such as --max-instances 25 --browser firefox --headless false to customize instance limits, browser type, headless mode, viewport size, and more.

Key features of Concurrent Browser MCP

Multi-instance concurrency with full resource isolation
Dynamic instance creation, management, and cleanup
Supports Chromium, Firefox, and WebKit
Automatic timeout-based cleanup to prevent leaks
Complete browser automation (navigation, click, type, screenshots)
Customizable viewport, user agent, and CSP bypass

Use cases of Concurrent Browser MCP

Parallel web scraping of multiple pages or sites
Running multiple automated test sessions simultaneously
Monitoring several web dashboards in real time
Concurrent form submissions and data extraction tasks
Multi-tab research and content gathering for AI assistants

FAQ from Concurrent Browser MCP

How does Concurrent Browser MCP differ from traditional MCP browser servers?

Traditional servers manage a single browser instance and execute requests serially, while Concurrent Browser MCP supports multiple isolated instances running in parallel, with automatic lifecycle management and cleanup for improved scalability and resource control.

What browsers are supported?

Chromium, Firefox, and WebKit are all supported. The default browser is Chromium, but you can switch using the --browser option.

How do I install and configure it?

Install via npm install -g concurrent-browser-mcp or use npx concurrent-browser-mcp. For MCP client integration, add a configuration entry under mcpServers with the command npx and args ["concurrent-browser-mcp", "--max-instances", "20"]. See the README for variants using global install, local build, or npm link.

What are the default settings?

Default maximum instances: 20; instance timeout: 30 minutes; cleanup interval: 5 minutes; headless mode: true; viewport: 1280×720. These can all be changed via command-line options.

How is resource cleanup handled?

The server automatically cleans up browser instances that exceed the configured timeout, preventing resource leaks. The cleanup interval (default 5 minutes) controls how often stale instances are removed.