Connect Your AI to STR Data in 30 Seconds

Claude Code is Anthropic's terminal-based AI coding agent and the most widely used AI dev tool of 2026. It supports MCP natively with a single CLI command -- no config files, no manual editing. This is the fastest path to connecting AirROI.

Step 1: Open your terminal

Launch your terminal application (Terminal on macOS, iTerm, Windows Terminal, or any shell). Claude Code must already be installed -- if not, run npm install -g @anthropic-ai/claude-code first.

Step 2: Run the setup command

This single command registers AirROI as an MCP server with HTTP transport and authenticates using your API key:

claude mcp add --transport http airroi https://mcp.airroi.com \
  --header "X-API-KEY: YOUR_API_KEY"

By default the server is added for the current project only. Add --scope user to make AirROI available in all your projects.

Step 3: Verify the connection

Start a new Claude Code session and type /mcp to check the server status, or ask: “What MCP tools do you have from AirROI?” You can also verify with claude mcp list.

Replace YOUR_API_KEY with your actual API key from the Developer Dashboard.

Claude Desktop is Anthropic's desktop app for macOS and Windows -- the easiest way to query AirROI data without touching a terminal after setup. Custom connectors are available on every plan, including Free (limited to one custom connector). The setup below uses the mcp-remote bridge, which requires Node.js.

Step 1: Install Node.js

Download and install Node.js from nodejs.org (any current version). Claude Desktop uses it to run the small bridge that connects to AirROI. If you already have Node.js, skip this step.

Step 2: Open the config file

In Claude Desktop, open Settings → Developer → Edit Config. This opens claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\).

Step 3: Add the configuration

Paste the following (or merge the airroi entry into your existing mcpServers block):

{
  "mcpServers": {
    "airroi": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.airroi.com",
        "--header",
        "X-API-KEY:YOUR_API_KEY"
      ]
    }
  }
}

Step 4: Restart and verify

Fully quit Claude Desktop (Cmd+Q on macOS, or quit from the system tray on Windows -- closing the window is not enough) and reopen it. Click the search-and-tools icon in the chat composer: the airroi tools should be listed. Then ask: “What AirROI tools do you have?”

Replace YOUR_API_KEY with your actual API key from the Developer Dashboard.

Codex is OpenAI's AI coding agent. It lives in three places -- the ChatGPT desktop app (Codex is built in since July 2026), the Codex CLI, and the Codex IDE extension for VS Code -- and all three read the same config file. This one edit connects them all.

Step 1: Open the config file

Codex stores MCP server configuration in TOML format at ~/.codex/config.toml (user-level) or .codex/config.toml (project-level, trusted projects only). Create the file if it doesn't exist.

Step 2: Add the MCP server configuration

Add the following TOML block. Codex uses http_headers for static header values (the API key is included directly):

[mcp_servers.airroi]
url = "https://mcp.airroi.com"
http_headers = { "X-API-KEY" = "YOUR_API_KEY" }

The desktop app and IDE extension also have a Settings → MCP servers → Add server dialog, but it has no field for custom headers -- the config file is the reliable path for API-key auth.

Step 3: Restart and verify

Restart the ChatGPT app or IDE extension (or start a new CLI session), then type /mcp in the Codex composer or run codex mcp list in your terminal to confirm AirROI appears.

Replace YOUR_API_KEY with your actual API key from the Developer Dashboard.

Cursor is an AI-first code editor with built-in MCP support. Ideal for developers who want to query STR data alongside their codebase -- building investment analysis scripts and pulling live market data within the same session.

Step 1: Open your MCP config file

Open Customize in Cursor's sidebar to manage MCP servers, or edit the config file directly at ~/.cursor/mcp.json (user-level) or .cursor/mcp.json (project-level).

Step 2: Add the configuration

Add the following to your config file. Cursor infers the transport from the URL -- no type field needed:

{
  "mcpServers": {
    "airroi": {
      "url": "https://mcp.airroi.com",
      "headers": {
        "X-API-KEY": "YOUR_API_KEY"
      }
    }
  }
}

Prefer keeping the key out of the file? Cursor resolves ${env:AIRROI_API_KEY} inside headers from your environment.

Step 3: Verify

Open a new Cursor chat and ask: “What AirROI tools do you have available?” Cursor should discover all 25 tools and confirm the connection.

Replace YOUR_API_KEY with your actual API key from the Developer Dashboard.

VS Code supports MCP servers through GitHub Copilot -- generally available since VS Code 1.102. Query STR data directly from your editor's Copilot chat.

Step 1: Create the MCP config file

Create a .vscode/mcp.json file in your project root (workspace-level), or open the Command Palette and run “MCP: Open User Configuration” for user-level setup.

Step 2: Add the configuration

Note: VS Code uses "servers" as the root key (not "mcpServers"):

{
  "servers": {
    "airroi": {
      "type": "http",
      "url": "https://mcp.airroi.com",
      "headers": {
        "X-API-KEY": "YOUR_API_KEY"
      }
    }
  }
}

Prefer not to store the key in the file? VS Code's inputs feature can prompt for it securely on first use -- see VS Code's MCP docs.

Step 3: Verify

Open Copilot Chat in VS Code and ask: “What AirROI tools are available?” Copilot should discover and list all 25 tools. You can also run “MCP: List Servers” from the Command Palette.

Replace YOUR_API_KEY with your actual API key from the Developer Dashboard.

AirROI's MCP server uses standard streamable HTTP transport, so any MCP client that supports custom headers can connect. Below are setup instructions for Windsurf (now Devin Desktop), Gemini CLI, and a generic fallback for any other MCP client.

Windsurf / Devin Desktop

Windsurf was renamed Devin Desktop in June 2026; the MCP config is unchanged. Edit ~/.codeium/windsurf/mcp_config.json or use the MCP Servers section in settings. Note: it uses serverUrl instead of url:

{
  "mcpServers": {
    "airroi": {
      "serverUrl": "https://mcp.airroi.com",
      "headers": {
        "X-API-KEY": "YOUR_API_KEY"
      }
    }
  }
}

Gemini CLI

If you still have access, one command connects AirROI:

gemini mcp add --transport http \
  --header "X-API-KEY: YOUR_API_KEY" airroi https://mcp.airroi.com

ChatGPT

ChatGPT's chat-side connectors (developer mode) currently support only OAuth or no-auth servers, so they can't send an API-key header. Use Codex instead -- it's built into the same ChatGPT desktop app and fully supports AirROI. See the Codex tab.

Any Other MCP Client

Provide these details to your client's MCP settings:

Transport: HTTP (Streamable HTTP)
URL: https://mcp.airroi.com
Headers:
  X-API-KEY: YOUR_API_KEY

The server is fully remote -- no local SDKs or processes required. Refer to your client's documentation for the exact steps to add an HTTP-based MCP server.

Replace YOUR_API_KEY with your actual API key from the Developer Dashboard.