google-search-console

This MCP server provides a programmable interface to the Google Search Console API. It exposes tools to list accessible sites, fetch search analytics data (with various filters and dimensions), inspect URLs in the index, and retrieve sitemaps for a site. You can invoke these capabilities from Claude Desktop or any client that speaks the MCP protocol, enabling conversational access to performance metrics and site health data from Google Search Console directly within chat.

To use it, configure the MCP server in your Claude Desktop setup (or via your MCP client) so that the google-search-console server points to this Python module (via -m main). Once running, you can issue natural-language prompts like: “List all sites I have access to,” “Show me the last 30 days of search impressions for example.com by query,” or “Inspect the URL https://example.com/page.” The server handles translating those requests into the appropriate Google API calls and returns structured results suitable for chat display and further reasoning.

Prerequisites

• Python 3.11 or higher
• pip (comes with Python)
• Virtual environment tool (venv recommended)
• Google Cloud project with the Google Search Console API enabled
• OAuth 2.0 credentials downloaded as credentials.json placed in the project root

Setup steps

1. Clone the repository

git clone https://github.com/metehan777/google-search-console-mcp.git
cd google-search-console-mcp

2. Create and activate a virtual environment

python3 -m venv fresh_env
# macOS/Linux
source fresh_env/bin/activate
# Windows
# .\fresh_env\Scripts\activate

3. Install dependencies in editable mode

pip install -e .

4. Prepare Google Cloud credentials

• Ensure credentials.json is in the project root (search-console-mcp/credentials.json)
• Do not commit credentials.json to version control (add to .gitignore)

5. Run the MCP server

python -m main

Note: For Claude Desktop integration, configure the mcpServers entry with the absolute path to your Python interpreter or virtual environment as shown in the README. The server runs in standard I/O mode for development/testing.

Tips and common issues:

• Ensure credentials.json is present in the repository root and that token.json is not committed.
• You may need to enable the Google Search Console API in your Google Cloud project and configure OAuth consent screen before first authentication.
• When testing with Claude Desktop, provide the absolute path to the Python interpreter inside your virtual environment in the mcpServers configuration, as shown in examples.
• If you encounter API quota or authorization errors, verify that the Google account used has access to the targeted Search Console properties and that the OAuth scopes include the required permissions for site, analytics, and indexing operations.