Cathedral — Persistent Memory for AI Agents

Cathedral — Persistent Memory for AI Agents is a remote MCP server: there is nothing to install. Its endpoint is https://cathedral-ai.com/mcp, served over Streamable HTTP. In an assistant that accepts remote MCP servers (often under a setting named connectors, integrations or tools), add a new server and give it this URL; in a client configured by file, add it as a remote (HTTP) server with the same URL.

The registry entry declares an HTTP header for the connection: Authorization (secret). A secret header usually carries an API key or token issued by the service; clients that accept custom headers set it with the server's URL.

It can also run locally from a Python package published to PyPI: cathedral-mcp version 1.1.0. It speaks MCP over stdio, so the client starts it as a program and talks to it through standard input and output. It needs Python; clients usually start it with uvx (from uv) or after pip install — the usual command is uvx cathedral-mcp. It reads these environment variables: CATHEDRAL_API_KEY (required, secret), CATHEDRAL_BASE_URL and CATHEDRAL_SANITISE; set them in the client's configuration for this server.

In the mcpServers JSON format that many desktop and editor MCP clients read, the entry looks like this (placeholders in angle brackets):

{
  "mcpServers": {
    "cathedral-mcp": {
      "command": "uvx",
      "args": [
        "cathedral-mcp"
      ],
      "env": {
        "CATHEDRAL_API_KEY": "<secret>",
        "CATHEDRAL_BASE_URL": "<value>",
        "CATHEDRAL_SANITISE": "<value>"
      }
    }
  }
}

Derived from the registry entry, not tested here. What the server does, and on what terms, is set by its publisher; check its repository or website before giving it access to your accounts or files. How to add an MCP server to an assistant · Before you connect