Connecting to the Selector MCP Server

Selector Logo

Overview

Selector supports Model Context Protocol (MCP) Server functionality to allow any MCP Client to interact with Selector to access supported APIs and tools. This guide provides step-by-step instructions for setting up MCP Clients to connect to the Selector’s MCP Server.

The Model Context Protocol (MCP) establishes rules for how an AI agent can find, connect to, and use external tools – whether it’s querying a database or running a command. This lets AI models go beyond their training data, which makes the AI more flexible and aware of the world around them.

MCP classifies AI entities into three main categories:

  • Hosts: LLM applications that initiate connections.
  • Clients: Connectors within the host application.
  • Servers: Services that provide context and capabilities.

MCP uses standard JSON-RPC 2.0 messages to communicate between these three types of AI entities.

There are three major steps to setting up MCP for connectivity:

  1. Obtain a Selector API Key
  2. Configure the MCP Client
  3. Verify the MCP Connection

Prerequisites

Before beginning the MCP setup process, you must have the following prerequisites:

  • A functioning MCP client (for example, Claude Desktop, Cursor, VSCode, and so on)
  • Access to the Selector S2AP software. Selector MCP Server is built into Selector’s S2AP
  • A valid Selector API key with appropriate permissions

Step 1: Obtain Your Selector API Key

  • Log in to Selector S2AP

    1. Navigate to your Selector S2AP instance (<customer-tenant-name>.selector.ai)
    2. Sign in with your credentials

  • Access the API Keys Section of S2AP. To do so:

    1. Click on your profile icon in the top right corner
    2. Select My API keys from the dropdown menu

  • Create a New API Key for MCP

    1. Click on Add new button
    2. Fill in the following fields:
      • Name: Give your API key a descriptive name (such as “MCP Key”)
      • Active: Ensure the checkbox is checked to make the key active
      • Roles: Select a role from the dropdown list of valid entries
      • Key valid until: Set the expiration date (default is one month from creation)

  • Generate and Save the Key

    1. Click the Add API key button
    2. IMPORTANT: Copy and save the generated API key immediately. You won’t be able to see it again!
    3. The key will look something like: qanMtokenxikKeBr0F

Step 2: Configure MCP Client

  • Selector MCP Details:

    • Selector MCP is a Streamable HTTP Remote Server (not SSE)
    • URL: https://<customer-tenant-name>.selector.ai/docs/mcp/
    • Auth Header Required
    • Auth Header should follow format “Authorization”: “YOUR_API_KEY_HERE”

  • Client Specific Documentation

  • Some Helpful MCP Client Specific Tips: Please note that these helpful tips are offered here to help you adopt MCP technology faster. Please refer to vendor documentation for more specific details.

    • For Cursor, there is a menu in settings for MCPs (or edit the .cursor/mcp.json file)
    • For VSCode, there is a command palette entry for adding MCP servers (or also a .vscode/mcp.json file)
      • >MCP: Add Server
        NOTE: The > sign (Ctrl+Shift+P) is used to find any show or run commands in the VS Code command palette.
    • For Gemini CLI, use the settings.json file to add MCPs
      • The Selector MCP is very similar to the Github example in the linked documentation above. You will use the “httpUrl” and “headers” keys.
    • For Claude Code, you can do this with a command:
      • claude mcp add --transport http <name> <url> --header “Authorization: \ Bearer YOUR_API_KEY_HERE”
    • For Claude Desktop, in the settings, go to Developer and click Edit Config. This opens the claude_desktop_config.json file. Remote MCP with Bearer token is not technically supported natively. For this, we can use “mcp-remote” like so:
`{`  
  `"mcpServers": {`  
    `"selector": {`  
      `"command": "npx",`  
      `"args": [`  
        `"mcp-remote",`  
        `"https://<customer-tenant-name>.selector.ai/mcp/",`  
        `"--header",`  
        `"Authorization: Bearer YOUR_API_KEY_HERE"`  
      `],`  
      `"env": {}`  
    `}`  
  `}`  
`}
  • An example of the mcp.json file in Cursor or VSCode:
`{`  
 `"mcpServers": {`  
   `"Selector MCP": {`  
     `"url": "https://<customer-tenant-name>.selector.ai/mcp/",`  
     `"headers": {`  
       `"Authorization": "Bearer YOUR_API_KEY_HERE"`  
     `}`  
   `}`  
 `}`  
`}`

Important Notes for Step 2:

  • Replace YOUR_API_KEY_HERE with the API key you generated in Step 1
  • Replace <customer-tenant-name>.selector.ai with your actual Selector instance URL
  • Make sure the url ends with /mcp/ (the trailing / is necessary)
  • Ensure the JSON syntax is valid (that is uses proper quotes, commas, etc.)
  • Save and Start (Restart)
    1. Save the configuration file
    2. Start (Restart)
    3. In some cases, another restart may be required

Step 3: Verify the Connection

  • Check MCP Status

    • Look for the MCP tools indicator
    • You should see Selector tools available

  • Test Basic Functionality by interacting with Selector

    • “Show me available Selector commands”
    • “What Selector tools do you have access to?”

Current Available Tools in Selector MCP Server

  • echo
  • ask_selector
  • query_selector
  • get_image_for_nl_query
  • list_maintwindows
  • create_maintwindow
  • get_maintwindow_by_name
  • update_maintwindow
  • delete_maintwindow_by_name
  • get_selector_phrases

To get a detailed list of the tools, ask for the available tools from Selector MCP.