HeadlessOps
MCP Setup

Connect Your AI Client

Connect any MCP-compatible AI client to headlessops.ai in under 2 minutes — Claude Code, Claude Desktop, Cursor, Codex, Windsurf, VS Code Copilot, or any generic client.

Connect Your AI Client

HeadlessOps exposes a native MCP server at:

https://app.headlessops.ai/mcp

Pick your client below, paste the config snippet, restart, and authorize in your browser.


Quick Setup

Run the following command in your terminal:

claude mcp add headlessops --transport http https://app.headlessops.ai/mcp

That's it. Claude Code will prompt you to authorize in your browser on first use.

Or configure manually in ~/.claude/claude_desktop_config.json (macOS/Linux) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "headlessops": {
      "url": "https://app.headlessops.ai/mcp",
      "type": "http"
    }
  }
}

Restart Claude Code and authorize when prompted.

Full Claude Code setup guide →

Edit your Claude Desktop config file:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "headlessops": {
      "url": "https://app.headlessops.ai/mcp",
      "type": "http"
    }
  }
}
  1. Save the file
  2. Restart Claude Desktop
  3. Authorize in the browser when prompted

Full Claude Desktop setup guide →

Open Cursor Settings → Features → Model Context Protocol and add:

{
  "name": "HeadlessOps",
  "url": "https://app.headlessops.ai/mcp",
  "type": "http"
}
  1. Save settings
  2. Restart Cursor
  3. Authorize in the browser when prompted

Full Cursor setup guide →

Add to your codex.json MCP configuration:

{
  "mcpServers": {
    "headlessops": {
      "url": "https://app.headlessops.ai/mcp",
      "type": "http"
    }
  }
}

Or with the local proxy (required for folderPath support):

{
  "mcpServers": {
    "headlessops": {
      "command": "npx",
      "args": ["-y", "@headlessops-ai/mcp", "--mcp-url", "https://app.headlessops.ai/mcp"]
    }
  }
}

Full Codex setup guide →

Open Windsurf Settings → MCP and add a new server:

{
  "mcpServers": {
    "headlessops": {
      "url": "https://app.headlessops.ai/mcp",
      "type": "http"
    }
  }
}

Or with the local proxy:

{
  "mcpServers": {
    "headlessops": {
      "command": "headlessops-mcp",
      "args": ["--mcp-url", "https://app.headlessops.ai/mcp"]
    }
  }
}

Full Windsurf setup guide →

Add to your VS Code settings.json:

{
  "mcp": {
    "servers": {
      "headlessops": {
        "url": "https://app.headlessops.ai/mcp",
        "type": "http"
      }
    }
  }
}

Or with the local proxy:

{
  "mcp": {
    "servers": {
      "headlessops": {
        "type": "stdio",
        "command": "headlessops-mcp",
        "args": ["--mcp-url", "https://app.headlessops.ai/mcp"]
      }
    }
  }
}

Full VS Code Copilot setup guide →

Any MCP client that supports HTTP or stdio transport works.

HTTP transport (direct URL):

{
  "url": "https://app.headlessops.ai/mcp",
  "type": "http"
}

stdio transport (via local proxy):

npm install -g @headlessops-ai/mcp
{
  "command": "headlessops-mcp",
  "args": ["--mcp-url", "https://app.headlessops.ai/mcp"]
}

macOS Homebrew Node: use /opt/homebrew/bin/headlessops-mcp as the command if your client doesn't inherit shell PATH.


Authentication

All connections use OAuth 2.1 with PKCE — no API key required. OAuth is user-level: each user authenticates individually and receives their own access token scoped to their account. On first connection, your AI client opens a browser window where you log in to HeadlessOps and grant access. The token is stored automatically.

This means multiple users can each connect their own AI client independently — every user authorizes with their own HeadlessOps account.


Working with Local Files (folderPath)

To let your AI agent read and write local files directly, install the local proxy:

npm install -g @headlessops-ai/mcp

Then replace the url config with:

{
  "command": "headlessops-mcp",
  "args": ["--mcp-url", "https://app.headlessops.ai/mcp"]
}

Next Steps