gsc

This MCP server integrates Google Search Console (GSC) data into Claude Desktop, enabling you to run analytical queries over your site’s search performance. Once running, you’ll have access to a set of GSC analytics tools that help you compare time periods, visualize trends with charts, and identify optimization opportunities based on metrics like clicks, impressions, click-through rate, and ranking shifts. The server relies on a Google service account credential file (JSON) to authenticate with Google APIs, so ensure you’ve generated and provided the correct credentials. In Claude Desktop, you’ll see five new tools associated with the MCP server that you can invoke after granting permission. Use prompts like asking for a date-range based analysis or a comparison between two periods to have Claude fetch and visualize the data through the MCP server.

Prerequisites:

• Python installed (and pip)
• uv (Python project manager)
• Claude Desktop installed
• Git (optional but recommended for cloning)

Steps:

1. Clone the repository

git clone https://github.com/seotesting-com/gsc-mcp-server.git
cd gsc-mcp-server

2. Create and activate a virtual environment using uv

# Windows
uv venv
.venv\Scripts\activate

# macOS/Linux
uv venv
source .venv/bin/activate

3. Install MCP server dependencies (and any package management steps if required by uv)

uv sync

4. Prepare your Google credentials and the environment variable for the MCP server

• Generate or obtain the service account JSON file from Google Cloud Console
• Note the full path to the JSON credentials file

5. Install the MCP server into Claude Desktop

mcp install server.py -v GOOGLE_APPLICATION_CREDENTIALS=<path-to-credentials-file>

Replace <path-to-credentials-file> with the actual path, e.g., C:\Users\Me\Downloads\credentials.json or /home/user/credentials.json 6) Restart Claude Desktop if needed

Tips and notes:

• The MCP server requires Google service account credentials with appropriate permissions in Google Search Console. Make sure the service account is added to the GSC property with at least restricted access.
• Set the GOOGLE_APPLICATION_CREDENTIALS environment variable to the JSON key path before starting the MCP server in Claude.
• If Claude Desktop cannot detect the MCP server after installation, ensure the virtual environment is active and that the server.py file is accessible to the runtime. Restart Claude Desktop if necessary.
• The tools exposed by Claude after enabling the MCP server include various search analytics capabilities such as site-wide queries, time-range comparisons, and performance visualizations. You may need to grant permission in Claude when prompted to run a tool.
• If you encounter API quota or authentication issues, verify that the service account key is valid and that the enabled Google APIs match your use case (Google Search Console API).