GDD   AI Controlled Browser Farm logo

GDD AI Controlled Browser Farm

Controls isolated Chromium browsers through MCP for multi-device testing, interaction, inspection, and environment simulation.

Install GDD AI Controlled Browser Farm

Pick your client, then add it in one click or paste the config.

No confirmed install command yet

We only publish a config when the install details come from the project itself — its README, its docs, or a verified owner. We haven’t found those for GDD - AI-Controlled Browser Farm, and we’d rather show nothing than a guess you’d paste into your client.

Open setup steps on GitHub

Overview

GDD - AI-Controlled Browser Farm MCP server lets an AI client control multiple isolated Chromium instances through MCP over HTTP. It supports navigation, real input events, screenshots, page inspection, device emulation, geolocation, network conditions, and browser diagnostics. Use it to test web applications across several device profiles or automate browser workflows from a headless machine or local desktop.

Use cases

  • Test signup and form flows across emulated devices
  • Run browser checks in parallel across isolated sessions
  • Inspect console errors and network activity during UI tests
  • Capture screenshots after automated interactions
  • Simulate location, language, and network conditions

Key features

  • Multi-instance isolated Chromium control
  • 22 device presets and custom viewports
  • Realistic mouse, touch, keyboard, and drag events
  • Geolocation, locale, language, and network emulation
  • Screenshots, console, network, and notification inspection
  • HTTP MCP endpoint with optional headless operation

Tools39

What GDD AI Controlled Browser Farm lets your agent do. Expand a tool for its parameters.

Self-reported~647 tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 3d ago

Recent health check succeeded.

9/9 checks healthy over the last 45d

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars12Stargazers on the source repository.
Last commit25d agoMost recent push to the default branch.
Tools exposed39Callable tools this server registers over MCP.

How GDD AI Controlled Browser Farm works

What GDD - AI-Controlled Browser Farm does

GDD - AI-Controlled Browser Farm MCP server exposes an HTTP MCP endpoint for operating multiple isolated Chromium instances. Each browser instance has its own profile, cookies, device settings, geolocation, and network conditions. Agents can create and remove browser instances, address each one by player ID, and inspect the current collection with its active state.

The server is suited to browser-based application testing and scripted interaction. An agent can open a site, complete forms, trigger touch or mouse interactions, capture screenshots, and inspect page output. Multiple browsers can be configured with different device presets so the same workflow can be checked across phones, tablets, and desktops.

How it works

GDD starts a headless or headed browser-control service and listens on port 9700 by default. MCP clients connect to http://localhost:9700/mcp. The agent first creates one or more players, optionally selecting a device preset, then passes the relevant player ID to navigation, input, reading, and diagnostic tools.

Input actions use browser-level events. Tapping chooses touch or mouse input according to the emulated device, while typing sends trusted key events and supports contenteditable elements. Optional humanization adds cursor paths or per-key timing variation. Dragging uses real pointer events and can target coordinates or another element. Page state can be checked through text reads, screenshots, console output, network details, notifications, and the current URL or title.

Setup and configuration

Download the server package for Windows, Linux, or macOS, or run the published Docker image. The server can also be installed for Claude Desktop through the provided .mcpb package. Linux, macOS, and Windows packages include platform-specific launch instructions; Chromium dependencies are included in the Docker image.

Start the binary directly, or add --headless for CI/CD environments. Other documented flags include --stealth, --stealth-max, --update, --version, and --help. Connect an MCP client with an HTTP server entry pointing to http://localhost:9700/mcp. Client configuration changes require restarting the client session.

GDD - AI-Controlled Browser Farm can alternatively be launched through the repository's Windows, Linux, or macOS stdio proxy scripts. The desktop edition is separate from the headless server and provides a browser-thumbnail grid for manually taking over sessions.

Tools and capabilities

The listed tools cover:

  • Creating, removing, listing, and inspecting browser instances.
  • Navigating, reloading, going backward or forward, and waiting for CSS selectors.
  • Tapping, swiping, dragging, scrolling, hovering, typing, pressing keys, selecting options, and handling JavaScript dialogs.
  • Reading one or many matching elements and capturing JPEG screenshots.
  • Applying device presets, viewport dimensions, location, timezone, locale, language, and network conditions.
  • Reading console output, uncaught exceptions, network requests, request timing, status codes, and push notifications.
  • Rewriting response headers to remove framing restrictions such as X-Frame-Options and CSP.

Limitations and notes

The server uses port 9700, while the optional desktop application uses port 9800. By default, browsers launch in headed mode, so headless operation must be selected for display-free environments.

Claude Code asks for confirmation on each MCP tool call by default; its documented permission setting can allow the GDD tool namespace without repeated prompts. The material does not specify authentication or remote-access controls for the HTTP endpoint, so deployment beyond the local machine should be evaluated separately.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about GDD AI Controlled Browser Farm

GDD - AI-Controlled Browser Farm is an MCP server that connects AI clients to isolated Chromium browser instances. Its main tools create and manage browsers, navigate pages, perform input actions, capture screenshots, inspect page content, and retrieve console or network diagnostics through an HTTP endpoint.

Install it by downloading the server package for your operating system, running the Docker image, or opening the Claude Desktop `.mcpb` package. Start the server, then configure your MCP client with `http://localhost:9700/mcp`.

Yes, Cursor is documented as a supported client. Add an HTTP MCP server entry pointing to `http://localhost:9700/mcp` in the project or global Cursor MCP configuration, then restart the client session.

Run the server with the `--headless` flag. Headless mode is documented for CI/CD and display-free environments; the Docker image also runs in headless mode by default.

Claude Desktop, Claude Code, Cursor, VS Code, and Windsurf are documented. Claude Desktop can use the provided `.mcpb` package, while the other documented configurations connect to the local HTTP MCP endpoint.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewGDD   AI Controlled Browser Farm AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/gdd-ai-controlled-browser-farm?style=directory)](https://allmcps.com/mcp/gdd-ai-controlled-browser-farm)
HTML Embed
<a href="https://allmcps.com/mcp/gdd-ai-controlled-browser-farm"><img src="https://allmcps.com/api/badge/gdd-ai-controlled-browser-farm?style=directory" alt="GDD AI Controlled Browser Farm on AllMCPs" /></a>