I built an open-source MCP server connecting Claude and Codex to Google Search Console, Bing and GA4
I built `Search Console MCP` (search-console.bruniaux.com), a Python server that lets Claude and Codex query your search and analytics data, then inspect the pages behind it. It works with other MCP clients too.
It connects Google Search Console, Bing Webmaster Tools, GA4 and Chrome UX Report, alongside technical page audits.
Some prompts you can use:
- “Find Google queries ranking 8–15, then inspect the relevant pages' headings and internal links.”
- “Compare Google and Bing performance, and flag anything that isn't directly comparable.”
- “Compare GSC clicks with GA4 organic sessions and highlight discrepancies.”
Google and Bing don't report positions the same way, so the comparison keeps them separate. It also omits cross-engine deltas unless the observed date windows are exact and equal, and reports unavailable metrics as unavailable.
Version `1.2.0` includes 81 tools, 19 of them for Bing. You can also use a CLI backed by the same registry.
The server runs locally and uses your own provider credentials. Google access requires OAuth or a service account; Bing requires a Webmaster API key. Setup is documented in English and French.
It's free and MIT licensed.
--> GitHub repository
--> Installation guide
PS: I'm also an RTK core team member and the author of the Claude Code Ultimate Guide. More about my work on my GitHub profile and my website.
If you use AI for SEO work, what task would you want to try with this first?
3 points · 8 comments
Mentions
I built an open-source MCP server connecting Claude and Codex to Google Search Console, Bing and GA4 I built `Search Console MCP` (search-console.bruniaux.com), a Python server that lets Claude and Codex query your search and analytics data, then inspect the pages behind it. It works with other MCP clients too. It connects Google Search Console, Bing Webmaster Tools, GA4 and Chrome UX Report, alongside technical page audits. Some prompts you can use: - “Find Google queries ranking 8–15, then inspect the relevant pages' headings and internal links.” - “Compare Google and Bing performance, and flag anything that isn't directly comparable.” - “Compare GSC clicks with GA4 organic sessions and highlight discrepancies.” Google and Bing don't report positions the same way, so the comparison keeps them separate. It also omits cross-engine deltas unless the observed date windows are exact and equal, and reports unavailable metrics as unavailable. Version `1.2.0` includes 81 tools, 19 of them for Bing. You can also use a CLI backed by the same registry. The server runs locally and uses your own provider credentials. Google access requires OAuth or a service account; Bing requires a Webmaster API key. Setup is documented in English and French. It's free and MIT licensed. --> GitHub repository --> Installation guide PS: I'm also an RTK core team member and the author of the Claude Code Ultimate Guide. More about my work on my GitHub profile and my website. If you use AI for SEO work, what task would you want to try with this first?