Google Search Console MCP Server
About
Google Search Console MCP Server for Claude, Cursor, Windsurf and other MCP Clients
Details
- Author
- surendranb
- GitHub stars
- 34
- Downloads
- 362
- Categories
- Search
Jump to
- Audits search appearance for cannibalization insights
- Segments traffic into "Searchers" and "Prompters"
- Recommends click triggers from high-visibility content
- Cross-checks pages with URL Inspection API for crawl errors
- Measures brand reference value vs. destination value
- Calculates intent efficiency for each search intent type
Install via pip install google-search-console-mcp. Configure your MCP-compliant AI agent by adding a JSON entry to its settings file with the command gsc-mcp and environment variables GOOGLE_APPLICATION_CREDENTIALS (path to your service account JSON key) and GSC_SITE_URL (e.g., sc-domain:example.com). Beforehand, you must set up a Google Cloud service account, enable the Search Console API, grant that account Full permissions in Google Search Console, and identify your property URL.
Google Search Console MCP "Intel Engine" ๐
The Authority-Based Visibility Governance Tool for the Evolving Search Landscape.
This is not just a data wrapper. It is a strategic "Intel" engine that transforms raw Google Search Console signals into actionable marketing insights. It is designed for marketers who need to understand their performance in a search landscape increasingly defined by AI Overviews and conversational search. Compatible with any MCP-compliant AI Agent.
๐ฏ Authoritative "Intel" Tools
| Tool Name | Actionable Marketing Intel Provided |
| :--- | :--- |
| get_search_appearance_audit | Cannibalization Intel. Detects if you are being used as a "Silent Reference" (high visibility but no clicks) in specialized SERP features. |
| get_intent_segmentation | Strategic Audience Intel. Segments traffic into "Searchers" (Traditional Keywords) vs. "Prompters" (Natural Language/AI Prompts). |
| identify_citation_opportunities | Growth Intel. Finds content that satisfies user intent so well that users don't click. Recommends "Click-Triggers." |
| get_technical_citation_audit | Technical Health Overlay. Cross-checks high-visibility pages with the URL Inspection API to find disqualifying crawl errors. |
| get_brand_visibility_summary | Brand Health Intel. Measures your Brand's "Reference Value" vs its "Destination Value." |
| calculate_intent_efficiency | Conversion Intel. Shows which search intent (Informational/Navigational) is most effectively driving site visits. |
---
๐ Getting Started
1. Google Search Console Setup
Before installing the MCP server, you must configure Google Cloud and Search Console access:A. Create Service Account:
1. Go to the Google Cloud Console.
2. Create a new project and enable the Google Search Console API.
3. Go to APIs & Services > Credentials and create a Service Account.
4. Create a JSON Key for the service account and download it (save as gsc-key.json).
B. Grant Access in Search Console:
1. Open your JSON key file and copy the client_email address.
2. Go to Google Search Console.
3. Select your property and go to Settings > Users and Permissions.
4. Click Add User, paste the service account email, and select Full permissions.
C. Identify Your Property URL:
- For Domain properties, use the format: sc-domain:example.com
- For URL-prefix properties, use the full URL: https://example.com/
2. Installation
pip install google-search-console-mcp
3. Configuration (Universal AI Agent)
Add this to your agent's MCP settings file:{
"mcpServers": {
"gsc-search": {
"command": "gsc-mcp",
"env": {
"GOOGLE_APPLICATION_CREDENTIALS": "/path/to/your/gsc-key.json",
"GSC_SITE_URL": "sc-domain:example.com"
}
}
}
}
---
๐ ๏ธ Project Philosophy
This project focuses on high-leverage data analysis for modern search: - Simplicity First: Minimum code for maximum insight. - Token Efficiency: Server-side aggregation prevents "Context Length" issues. - Authoritative Data: We only use official Google Search Console API signals. No speculative "AI SEO" hacks.---
License
MIT LicenseSign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.

