Define the release window
Supply the deployment date, affected property and exact URL set. Compare complete, same-length pre/post periods and allow for GSC reporting delay.
Google Search Console + AI
Investigate a release from the terminal with both your repository and search performance in view. AskGSC supplies read-only GSC evidence through MCP; Claude Code can inspect code and propose changes through its own development environment.
14 days free · No credit card
Read-only access
Correlating a release date with falling clicks is easy; attributing causality is harder. Claude Code can compare pre- and post-release Search Console windows, identify affected query–page pairs, then inspect the actual files changed in your repository. The investigation should rule out demand shifts before proposing a patch.
Supply the deployment date, affected property and exact URL set. Compare complete, same-length pre/post periods and allow for GSC reporting delay.
Check pages and queries that moved, together with impressions, clicks, CTR and average position. Ask whether unrelated pages show similar demand changes.
Use Claude Code's repository context to examine relevant commits, templates or metadata. Request a proposed patch only if the evidence points to an actionable issue; review before applying.
These prompts separate evidence from source control. GSC has no knowledge of git commits; Claude Code must obtain those from your repository.
Use the exact siteUrl returned by the property-list tool. Do not assume missing query rows are zero.
Explore 18 more Search Console promptsWith AskGSC, compare [PROPERTY] for the 28 complete days before [RELEASE_DATE] versus 28 complete days after it. Identify the URLs and queries with the biggest click changes. Then inspect git changes in this repo around the release. Label correlation versus verified causes and do not edit yet.
For [PROPERTY], compare traffic for URLs matching [PAGE_PREFIX] across two equal complete periods around [RELEASE_DATE]. Show affected pages by clicks, impressions and CTR. Inspect the matching template changes in the repository and tell me whether the evidence suggests a shared problem or isolated page variation.
For [PROPERTY] and [PAGE URL], compare two complete 28-day periods and break down changes by query. Review related recent commits and SEO metadata in Claude Code. Suggest the smallest reversible patch, explain your uncertainty and specify how to validate it after deployment.
Several URLs from one template lose clicks while other sections stay roughly stable.
A relevant title, canonical or content-template change exists in the release diff; this is separate from GSC data.
Reproduce and review the problem, apply a reversible fix if warranted, then compare full reporting windows. A matching date alone is not evidence of causation.
AskGSC can list accessible properties, query Search Analytics and compare periods. It cannot read git history, detect technical SEO defects, run deployment checks, inspect indexing or change Search Console settings. Claude Code's code access and MCP server setup are handled separately by the client.
A practical look at capabilities, reporting boundaries and the MCP connection.
Claude Code supports MCP servers. AskGSC supplies a hosted read-only GSC endpoint to compatible clients after authorization.
No. It exposes historical search metrics, not deploy logs or causality. Claude Code can compare those metrics with the repository and flag hypotheses.
AskGSC never modifies files. Claude Code may propose changes with its own development tools, but those require review and validation.
Use comparable GSC periods and your repository history to decide whether a release deserves investigation.