This panel lets you configure and monitor the built-in MCP Server. PRIVACY CAUTION: when using the MCP Server, your AI desktop client has access to all the data in your databases. That data is sent to the AI service for processing.
Enable MCP Server – check this box to enable the built-in MCP Server. The Search Warehouse application must be running for the MCP Server to work. It may stop working if left running for an extended period of time.
Port – by default, the MCP Server uses port 5328. If this conflicts with another service you are running, you can change it.
Name / URL – the name and URL of the MCP Server are listed to aid in configuring your AI desktop client. The MCP Server uses streamable HTTP. The configuration varies depending on which AI client application you are runnning.
MCP Server Session Log – while the application if open, a session log is running to capture the calls/responses to the MCP Server. If the response contained ‘SQL’, you can click on the entry to see/execute the SQL query and see what data was delivered to the AI client.
SQL Queries Library – if you want to save any of the queries, click on the entry to load/run it, then click on a Save New SQL Query button to save it to the library for that specific client or as a template for any client.
Sample AI client configurations — note that the examples below show you what worked for me with the installation I used. Check the documentation for your AI client application to see how to configure your installation.
For Claude Desktop: only supports stdio by default, so you need to use npx with mcp-remote
"mcpServers": {
"analyticsedge-search-console-desktop": {
"command": "npx",
"args": [
"-y",
"mcp-remote@latest",
"http://localhost:5328/mcp"
]
}
}
For VS Code/Cline:
"mcpServers": {
"analyticsedge-search-console-desktop": {
"type": "streamableHttp",
"url": "http://localhost:5328/mcp"
}
}
For Antigravity:
"mcpServers": {
"analyticsedge-search-console-desktop": {
"serverUrl": "http://localhost:5328/mcp"
}
}