"Which queries lost the most clicks last month?"
Ranked by click delta, with impressions and average position beside each one so a ranking drop is distinguishable from a demand drop.
Search Console MCP
Google Search ConsoleMCP for AI agents
Connect Search Console to your agent so it can answer with your clicks, impressions, CTR, and rankings. OpenRush adds live market context around those results.
Find what changed
The queries and pages that moved, without building a single report.
"Which pages gained impressions but lost click-through rate?"
The classic signal that a title or description is underperforming a ranking you already earned.
"Where did my average position slip week over week?"
Movement by query and by page, so the agent can point at a cause instead of a chart.
Close the loop with market data
Your own performance is only half the picture. OpenRush already holds the other half.
"I dropped on this keyword. Who took the position?"
Joins your performance against a live SERP reading, so the answer names the page that replaced you.
"Which of my ranking keywords have the most headroom?"
Your positions weighed against real search volume and CPC from the market index.
"What are competitors ranking for that I already have pages about?"
Keyword gap analysis narrowed to the pages you have actually published.
No Google Cloud setup
Most Search Console MCP servers are open source, which means you own the plumbing before you get an answer.
Running your own Search Console MCP server
- Create a Google Cloud project
- Enable the Search Console API
- Configure an OAuth consent screen
- Download and store a credentials file
- Run and keep a local server alive
Connecting through OpenRush
- Add the OpenRush MCP to your agent once
- Authorize your Search Console property from the dashboard
- Ask your question
Nothing to host, no credentials file on disk, and the same connection already serving your market data.
Connect Search ConsoleCommon Search Console MCP questions
A Google Search Console MCP is a Model Context Protocol server that connects your Search Console property to an AI agent, so the agent can query your clicks, impressions, click-through rate, and average position directly in conversation instead of you exporting a report and interpreting it yourself.
No. The open-source Search Console MCP servers require you to create a Google Cloud project, enable the Search Console API, configure an OAuth consent screen, and keep a credentials file on disk. OpenRush handles authorization for you, so connecting a property is a few clicks from your dashboard connections page.
Any MCP-compatible client, including Claude, ChatGPT, Claude Code, Codex, Cursor, and Perplexity. The connection is the same one-minute setup as the rest of OpenRush, so there is no second server to run.
Two things. It answers in conversation, so a question like which pages lost click-through rate does not become a filtering exercise. It also shares a connection with OpenRush's market data, so one answer can weigh your own numbers against live SERPs, competitor positions, and keyword demand in a single answer. Search Console alone cannot tell you who took the position you lost.
No. OpenRush reads a connected property only to answer your agent's questions, and never uses it for model training. See the privacy policy for how connected accounts are handled, and disconnect a property at any time from the dashboard.
Ask your search data a question.
One-click setup with Claude or any MCP-compatible agent.