Search Console

Provides read-only access to Google Search Console data for AI assistants to analyze site performance.

other

0

GitHub Stars

other

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "chrishart0-searchconsole-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/path/to/searchconsole-mcp",
        "searchconsole-mcp"
      ],
      "env": {
        "GOOGLE_APPLICATION_CREDENTIALS": "/path/to/key.json"
      }
    }
  }
}

This MCP server provides your AI assistants with read-only access to Google Search Console data. You can ask natural-language questions about your site’s search performance—keywords, clicks, impressions, CTR, rankings, sitemaps, and more—and receive structured insights without exposing credentials or allowing edits.

How to use

You connect your MCP client to the Search Console MCP server to start asking questions about your site's performance. Use a compatible MCP client that supports stdio connections. Once connected, you can request information such as top keywords by clicks, pages with high impressions but low CTR, your list of verified properties, and your sitemaps. The server answers in natural language, and can filter data by date ranges, dimensions, and search type.

How to install

Prerequisites you need before installing: Git, Python (for credentials tooling), and access to install or run the MCP client tooling (such as uv). Ensure you have internet access to clone repositories and install dependencies.

git clone https://github.com/chrishart0/searchconsole-mcp.git
cd searchconsole-mcp
uv sync

Additional content

Authentication uses Google Application Default Credentials. You may provide a service account key for automation or use user credentials for local development. The service account or user must have access to the Search Console properties you want to query.

Connecting your MCP client to the server requires configuration in your MCP client. The following two examples show how to configure popular MCP clients to run the server locally and connect via stdio.

For Claude Code users, add the following to your client configuration to launch the MCP server via UV. This configuration uses a local directory path to the server and injects the Google credentials path.

{
  "mcpServers": {
    "searchconsole_mcp": {
      "type": "stdio",
      "command": "uv",
      "args": ["run", "--directory", "/path/to/searchconsole-mcp", "searchconsole-mcp"],
      "env": {
        "GOOGLE_APPLICATION_CREDENTIALS": "/path/to/key.json"
      }
    }
  }
}

Examples of prompts to try

Once connected, try prompts such as: “What are my top 10 keywords by clicks this month?”, “Show me pages with high impressions but low CTR”, “List all my Search Console properties”, “What queries is my site ranking for on page 2 of Google?”, and “Show me my sitemaps and their status.”

Available tools

list_sites

Lists all verified Search Console properties with permission levels

query_search_analytics

Query keyword or page data such as clicks, impressions, CTR, and average position with support for filtering by dimension, date range, and pagination

list_sitemaps

Lists submitted sitemaps for a property