How to Connect Google Search Console to Cursor and Codex
Quick answer: Cursor reads Google Search Console through a remote MCP server. Add a server entry with the MCP URL to your global or project mcp.json, save, and complete the OAuth sign-in. In Codex, run codex mcp add with the same URL, then codex mcp login. Both editors can then query clicks, impressions, and sitemap status while you work.
You are about to rewrite the title tag and merge two pages, and the one question that matters lives in another tab: which queries actually send clicks to this URL? So you alt-tab to Search Console, rebuild the filter, lose the thread, and half the time ship the change without looking.
If you work in Cursor or Codex, the data can come to you. Both support remote MCP servers, so a read-only Search Console connection makes search performance a question you ask inline, mid-change, with the file still open.
SEARCH CONSOLE
How to Connect Google Search Console to Cursor and Codex
PRIMARY KEYWORD
connect Google Search Console to Cursor
READ TIME
8 minutes
Why connect Search Console to a code editor at all?
Because the riskiest SEO decisions happen during development, not during reporting. Deleting a page, changing a URL, rewriting titles, consolidating content: each is safer when you can see the queries and clicks attached to the URL before the commit.
An agent with Search Console access can also fold data into its work. Ask Cursor's agent to list your top landing pages before restructuring navigation, or to confirm a sitemap has no errors after a deploy, and it can check rather than assume.
How do you add a Search Console MCP server to Cursor?
Cursor reads MCP configuration from a global or per-project mcp.json file. Under the mcpServers key, add an entry named something like seo-receipts whose url field is the MCP server URL, for example the free SEO Receipts endpoint at https://seoreceipts.com/mcp. Save the file.
Cursor detects the server and prompts for OAuth sign-in. Approve it, choose which Search Console properties the editor may read, and the tools appear in the agent's tool list. A project-level file keeps client data scoped to the client's repo, which is usually the right call for consultants.
How do you add the same server to Codex?
Codex manages MCP servers from the command line. Run codex mcp add seo-receipts with the server URL passed as the url flag, then run codex mcp login seo-receipts. Your browser opens the authorization screen, where you pick the properties to share and return to the terminal.
After login, the Search Console tools are available in Codex sessions. The same read-only boundary applies: the agent can query performance rows and sitemap status, and cannot modify anything in your Google account.
What can you actually do with it during development?
The pattern that pays off is checking before changing. Every ask below is one prompt in the editor instead of a Search Console session in another window.
Keep the questions scoped to the URL or section you are touching. Editor agents inherit the same failure modes as any assistant, and a tight scope keeps the tool calls honest.
- Before deleting or merging a page, list the queries and clicks it earned in the last finalized 28 days
- Before rewriting a title, check the page's top queries and current CTR
- After changing routes, confirm the sitemap was re-read without errors
- Before a redesign, pull the top landing pages so nothing load-bearing gets dropped
- After a deploy, compare the section's clicks against the prior period
What limits should you expect from the connection?
The data limits are Search Console's. Final numbers arrive after a delay of a few days, history reaches back about 16 months, and anonymized queries stay hidden. Row limits and rate limits depend on the server; the SEO Receipts server returns up to 1,000 rows per query and caps request rates per account.
Editor agents also act autonomously, so a read-only server matters more here than anywhere. A misfired analysis query wastes a request. There is no write tool to misfire.
TAKEAWAY CHECK
What belongs on the wire search console into your editor safely?
- 01Choose a read-only Search Console MCP server.
- 02Add it to Cursor's mcp.json or via codex mcp add.
- 03Complete the OAuth sign-in and select only the needed properties.
- 04Test with a URL whose numbers you already know.
- 05Check query and click data before deleting, merging, or retitling pages.
- 06Keep client data in project-level config, and revoke unused grants.
What else do people ask about connect Google Search Console to Cursor?
Does this work with editors other than Cursor and Codex?
Any client that supports remote MCP servers can use the same URL, and MCP support is spreading across AI coding tools quickly. Check your editor's documentation for remote or streamable HTTP server support, then follow its config format with the same endpoint.
Can the editor's agent change my Search Console settings?
Not through a read-only server. The exposed tools query performance data, list properties, and read sitemap status. Submitting sitemaps, changing settings, or removing URLs would require write tools, which a read-only analysis server deliberately does not expose.
Should the MCP config be global or per project?
Per project when client data is involved, so a repo only sees the properties tied to that client. Global config is fine for a solo developer working on their own sites. Either way, the OAuth grant controls which properties are actually readable.
Why does the agent see no properties after connecting?
The Google account you authorized has no Search Console access, or none was selected during the grant. Verify the property in Search Console with that account, then redo the sign-in and select it explicitly when the property list appears.
Is editor access enough to replace Search Console itself?
No. The connection covers performance rows and sitemap status, which answers most day-to-day development questions. Index coverage details, manual actions, enhancement reports, and removals still live in the Search Console interface and deserve a regular direct look.
Which primary sources support this guide?
Product behavior and metric definitions change. These are the official Google references used for this article and checked on July 27, 2026.
- 01Cursor: Model Context Protocol
Cursor's documentation for configuring remote MCP servers in mcp.json.
- 02OpenAI: Codex MCP
OpenAI's instructions for adding and authorizing MCP servers in Codex.
- 03Model Context Protocol documentation
The open specification for connecting AI assistants to external tools and data sources.
- 04Search Analytics API query reference
Google's API parameters, authorization scopes, row limits, and incomplete-data metadata.
FREE GOOGLE SEARCH CONSOLE MCP SERVER
Ask your AI assistant about your own Search Console data.
Connect the free SEO Receipts MCP server once. Claude, ChatGPT, Cursor, or Codex can then read clicks, impressions, queries, pages, and sitemap status from the properties you select.