Skip to main content

Add an MCP Server

The Model Context Protocol (MCP) lets AI models use tools and data from other services. If a service offers an MCP server, you can connect it to ConsoleX by its URL and use its tools in chats, tasks and agents.

Add the server​

  1. Click … (More) in the left rail, then Capabilities.
  2. On the Tools tab, click Add custom tool, then MCP server: "Connect an existing MCP server by its URL."

Add custom tool menu with MCP server and HTTP API

  1. In the Add a MCP server dialog, fill in:
    • Name the MCP server: a name you'll recognize in tool lists.
    • Description of the server: what the server does.
    • Connection Type: Streamable HTTP or SSE, whichever your server supports. Its documentation tells you which.
    • Authentication, in the second drop-down: Bearer Token, OAuth or No Authentication.
    • HTTP URL: the server's address, for example https://example.com/mcp.
    • Bearer Token: shown when you choose Bearer Token. Paste the token the service gave you.
  2. Open Show advanced options only if your server's documentation asks for extra settings.
  3. Click Save.

Add a MCP server dialog with name, description, connection type, authentication, HTTP URL and Bearer Token

Use the server's tools​

Your MCP server is listed on the Tools tab under the Custom filter. Turn it on where you need it:

  • In a chat: click Tools under the message box and switch it on.
  • In a task or agent: click Add capabilities and switch it on.

If the server adds new tools later, they wait for your approval before they can be used.

Local (Stdio) servers​

ConsoleX is a cloud service, so it connects to MCP servers by URL over Streamable HTTP or SSE. An MCP server that only runs on your own computer over Stdio can't be added directly; it needs to be reachable at a URL first.