mcp-gsc-advanced
MCP server for advanced Google Search Console analysis — keyword cannibalization detection, page-level query deep dive, and rank change tracking.
README
mcp-gsc-advanced
Built by Artur Ferreira @ The GEO Lab · 𝕏 @TheGEO_Lab · LinkedIn · Reddit
MCP server for advanced Google Search Console analysis — keyword cannibalization detection, page-level query deep dive, and rank change tracking.
Extends basic GSC data into actionable SEO intelligence.
Tools
| Tool | Description |
|---|---|
detect_cannibalization |
Find queries where 2+ pages compete — severity scoring, impressions split, consolidation recommendations |
page_queries |
All queries driving traffic to a specific page — sorted by impressions with CTR and position |
rank_changes |
Period-over-period position changes — finds significant gains and losses |
Features
✅ GEO-native — built alongside the GEO Brand Citation Index, tracking brand visibility across ChatGPT, Perplexity, and Gemini
Install
# Claude Code (service account key file)
claude mcp add gsc-advanced \
-e GOOGLE_SERVICE_ACCOUNT_KEY=/path/to/service-account.json \
-e GSC_SITE_URL=https://yoursite.com/ \
-- npx mcp-gsc-advanced
# Or in .mcp.json
{
"mcpServers": {
"gsc-advanced": {
"command": "npx",
"args": ["mcp-gsc-advanced"],
"env": {
"GOOGLE_SERVICE_ACCOUNT_KEY": "/path/to/service-account.json",
"GSC_SITE_URL": "https://yoursite.com/"
}
}
}
}
Authentication
Requires a Google service account with Search Console access.
Option A: Set GOOGLE_SERVICE_ACCOUNT_KEY to the path of your service account JSON file.
Option B: Set GOOGLE_CLIENT_EMAIL and GOOGLE_PRIVATE_KEY env vars directly.
Then add the service account email as a Full user in Google Search Console.
Usage
> detect keyword cannibalization on my site
> what queries drive traffic to /geo-stack/ page?
> show rank changes in the last 7 days
Attributions & Licence
Built and maintained by Artur Ferreira @ The GEO Lab.
Email: artur@thegeolab.net
Best Practice Attribution
This MCP server was built following the open source Best Practice Approach — reading community work for inspiration, then writing original content, and crediting every source.
Based on:
- Model Context Protocol specification by Anthropic
- MCP SDK (MIT)
GSC analysis concepts inspired by:
- Google Search Console API — official documentation
- Aleyda Solís — GSC cannibalization detection methodology
- Kevin Indig / Growth Memo — GSC query analysis for AI search
Dependencies:
- googleapis — Google APIs Node.js client (Apache 2.0)
All server code is original writing. No files were copied or adapted from any source. MIT licence.
Found this useful? ⭐ Star the repo and connect: 🌐 thegeolab.net · 𝕏 @TheGEO_Lab · LinkedIn · Reddit
Related Repos
- claude-code-mcps — All 5 MCP servers in one collection
- mcp-seo-auditor — On-page SEO audit + JSON-LD validation
- mcp-serp-intel — SERP weak spots, PAA trees, intent comparison
- mcp-common-crawl — Free backlink discovery via Common Crawl
- mcp-gsc-advanced — GSC cannibalization, rank changes
- mcp-wordpress-setup — WordPress MCP server setup guide
Licence
MIT — see LICENSE
Built and maintained by Artur Ferreira @ The GEO Lab · MIT License
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。