Desktop Control

v1.0.0Browser Automationstable

Agentify Desktop lets Codex/Claude/OpenCode control your logged-in ChatGPT, Claude, AiStudio, Gemini, Grok, Perplexity web sessions via MCP, parallel hidden/visible tabs, file upload + image download

aistudiochatgptcodexelectrongemini
Share:
398
Stars
0
Downloads
0
Weekly
0/5

What is Desktop Control?

Desktop Control is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to agentify desktop lets codex/claude/opencode control your logged-in chatgpt, claude, aistudio, gemini, grok, perplexity web sessions via mcp, parallel hidden/visible tabs, file upload + image download

Agentify Desktop lets Codex/Claude/OpenCode control your logged-in ChatGPT, Claude, AiStudio, Gemini, Grok, Perplexity web sessions via MCP, parallel hidden/visible tabs, file upload + image download

This server falls under the Browser Automation category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • Agentify Desktop lets Codex/Claude/OpenCode control your log

Use Cases

Control your ChatGPT, Claude, Gemini, and Grok sessions through Codex and Claude.
Manage parallel browser tabs and upload files with AI automation.
agentify-sh

Maintainer

LicenseMPL-2.0
Languagejavascript
Versionv1.0.0
UpdatedMay 20, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx desktop

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Desktop Control

Agentify Desktop is an Electron-based MCP server that lets AI coding agents such as Claude Code, Codex, and OpenCode drive your already-logged-in ChatGPT, Claude, Gemini, Grok, Perplexity, and AI Studio web sessions through a unified MCP interface. It manages parallel hidden or visible browser tabs, handles file uploads, downloads generated images, and provides tools for querying AI assistants without sharing credentials or API keys. This makes it possible for one AI agent to orchestrate multiple other AI web sessions as part of an automated workflow.

Prerequisites

  • Node.js 20 or newer
  • An MCP-capable CLI client: Claude Code, Codex, or OpenCode
  • Active browser sessions for the AI platforms you want to control (ChatGPT, Gemini, etc.) — no API keys required
  • Electron-compatible operating system (macOS, Windows, or Linux)
1

Install Agentify Desktop

Run the package directly via npx (no global install required). This will download and launch the Electron app.

npx @agentify/desktop
2

Log in to your AI web sessions

Open ChatGPT, Claude, Gemini, Grok, or Perplexity in the Agentify Desktop browser window and log in normally. Agentify uses these existing sessions — no API keys are shared with the MCP layer.

3

Register as an MCP server in Claude Code

Add Agentify Desktop as an MCP server in your AI coding client. For Claude Code, run the following command.

claude mcp add --transport stdio agentify-desktop -- npx -y @agentify/desktop mcp
4

Or add to Claude Desktop config manually

Alternatively, add the server block to claude_desktop_config.json for use with Claude Desktop.

{
  "mcpServers": {
    "agentify-desktop": {
      "command": "npx",
      "args": ["-y", "@agentify/desktop", "mcp"],
      "env": {}
    }
  }
}
5

Show newly created tabs (optional)

Set the AGENTIFY_DESKTOP_SHOW_TABS environment variable or use the --show-tabs flag to make Agentify display each browser tab it creates, which helps with debugging.

npx @agentify/desktop mcp --show-tabs
6

Use agentify_ensure_ready before querying

In your agent workflow, call the agentify_ensure_ready tool before sending prompts to wait for the web session to be fully authenticated and the UI to be ready.

Desktop Control Examples

Client configuration

Add Agentify Desktop to claude_desktop_config.json to let Claude Desktop orchestrate other AI web sessions.

{
  "mcpServers": {
    "agentify-desktop": {
      "command": "npx",
      "args": ["-y", "@agentify/desktop", "mcp"],
      "env": {
        "AGENTIFY_DESKTOP_SHOW_TABS": "true"
      }
    }
  }
}

Prompts to try

Use these prompts with Claude Code or Codex after Agentify Desktop is connected as an MCP server.

- "Open a ChatGPT tab with key 'gpt-review' and ask it to summarize this repository in 5 bullets"
- "Send this code snippet to Gemini and ask for a security review"
- "Query Perplexity for the latest news on Model Context Protocol and return the top 3 results"
- "Upload this image to Claude web and ask it to describe what it sees"
- "Check the status of all open Agentify tabs"

Troubleshooting Desktop Control

Agentify Desktop cannot find existing browser sessions or shows a login page

By default Agentify uses isolated browser profiles. Set AGENTIFY_DESKTOP_CHROME_PROFILE_MODE=existing and AGENTIFY_DESKTOP_CHROME_PROFILE_NAME to your Chrome profile name to reuse an existing logged-in profile.

agentify_query times out waiting for the AI platform response

Always call agentify_ensure_ready for the tab key before agentify_query. If the session has expired, navigate to the login page using agentify_navigate and re-authenticate manually in the visible window.

npx @agentify/desktop fails on Linux with Electron errors

Electron requires a display server. On headless Linux, install Xvfb and run 'xvfb-run npx @agentify/desktop mcp'. Also ensure libnss3 and libatk1.0-0 system packages are installed.

Frequently Asked Questions about Desktop Control

What is Desktop Control?

Desktop Control is a Model Context Protocol (MCP) server that agentify desktop lets codex/claude/opencode control your logged-in chatgpt, claude, aistudio, gemini, grok, perplexity web sessions via mcp, parallel hidden/visible tabs, file upload + image download It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Desktop Control?

Follow the installation instructions on the Desktop Control GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Desktop Control?

Desktop Control works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Desktop Control free to use?

Yes, Desktop Control is open source and available under the MPL-2.0 license. You can use it freely in both personal and commercial projects.

Browse More Browser Automation MCP Servers

Explore all browser automation servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "desktop": { "command": "npx", "args": ["-y", "desktop"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Desktop Control?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides