How-To: Run the FastMCP Server

This guide explains how to configure, start, and debug the local Model Context Protocol (MCP) server to expose the Sovereign Palace to AI editors.

This guide explains how to configure, start, and debug the local Model Context Protocol (MCP) server to expose the Sovereign Palace to AI editors.

Prerequisites

• Python 3.11+ and uv must be installed.
• MCP client (e.g. Cursor, Claude Desktop, or Windsurf) running locally.

Step 1: Start the server

Launch the server using uv run to isolate and resolve dependencies dynamically. This starts the JSON-RPC interface over standard input/output (stdio):

uv run tools/mcp/server.py

Step 2: Configure AI editors (Cursor / Claude Desktop)

To connect your IDE to the server, append the following configuration block to your client settings.

Claude Desktop config (claudedesktopconfig.json)

Open ~/.code/claude_desktop_config.json (macOS/Linux) or %APPDATA%/Claude/claude_desktop_config.json (Windows) and insert:

{
  "mcpServers": {
    "dsom-palace": {
      "command": "uv",
      "args": [
        "run",
        "--with", "mcp[cli]>=1.1.2",
        "--with", "fastmcp>=0.1.0",
        "--with", "pyyaml>=6.0",
        "tools/mcp/server.py"
      ],
      "env": {
        "DSOM_ROOT": "/absolute/path/to/deep-state-of-mind-for-my-ai"
      }
    }
  }
}

Step 3: Verify server functionality

Execute the standard test suite to confirm that the server registers all required resources and tools:

uv run --with pyyaml --with pytest --with mcp==1.2.1 --with fastmcp pytest tests/test_mcp_server.py

---

Deep State of Mind (DSOM) For My AI Protocol | Harisfazillah Jamel (LinuxMalaysia) | 2026-08-14 Standard: UK English | DBP-standard Bahasa Melayu Malaysia (Piawai) | GNU General Public License v3.0