> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mapxagent.com/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP connections

> How to connect your own external tools to MapX through the Model Context Protocol (MCP).

MapX supports the **Model Context Protocol (MCP)**, a standard way to connect an AI model to external tools. You can connect your own MCP servers so the AI can call your internal tools during a conversation.

## Custom MCP servers

You can connect your own MCP servers so the AI can use your internal tools during a conversation:

1. Open **MCP connections** in the dashboard.
2. Add a server by URL (Streamable HTTP) or by local command.
3. Test the connection, then enable the tools you want to expose.

The AI loads the enabled tools into each conversation turn. Secrets such as bearer tokens or header values are stored encrypted and never returned by the API.

You can also disable individual tools. A disabled tool is never offered to the AI.

## Using MCP in chat

Connected tools work automatically. Ask for the outcome — "use my weather API to check conditions at each store" — and the AI calls the tool, reads the result, and continues.

<Note>
  MCP connections are optional. You do not need to set one up to use MapX's
  core map, layer, and analysis features.
</Note>

See [Agent tools](/en/integration/agent-tools) to connect external LLM agents to MapX instead.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.