Google Search Console

Query Google Search Console search analytics data. Returns clicks, impressions, CTR, and position. Args: site_url (string, required): The Search Console property URL. start_date (string, required): Start date in YYYY-MM-DD format. end_date (string, required): End date in YYYY-MM-DD format (must be >= start_date). dimensions (array, optional): Group by dimensions. Valid: query, page, device, country, date, searchAppearance. Note: searchAppearance cannot be combined with query, page, device, or country — only with date. search_type (string, optional): web (default), image, video, news, discover, googleNews. dimension_filter_groups (array, optional): Filter by dimension values. Note: Cannot combine aggregate_by=byPage with a page dimension filter. aggregate_by (string, optional): auto (default), byPage, byProperty, byNewsShowcasePanel. row_limit (int, optional): Rows per page. Range: [1, 25000]. Default: 25. cursor (string, optional): Opaque cursor from a previous response's pagination.nextCursor. Omit for the first page; never construct one by hand. Returns: data: array of { keys, clicks, impressions, ctr, position } rowCount: rows in this page (Search Console reports no grand total) pagination: { hasMore, nextCursor? } — nextCursor is absent on the last page responseAggregationType, firstIncompleteDate (if data is incomplete) Pagination: when pagination.hasMore is true, call again with the same parameters plus cursor set to pagination.nextCursor. A cursor only continues the exact call that returned it — changing the dates, dimensions or filters invalidates it, and the call is rejected rather than silently returning rows from a different query. IMPORTANT — when you need a complete, consistent result set, raise row_limit and take it in ONE call rather than crawling. Search Console sorts rows by clicks descending and orders ties "in an arbitrary way", and that tie order is not stable between calls. Since most rows on a typical property have 0 clicks, they are nearly all tied, so a multi-page crawl can repeat some rows and miss others. A single call has no such problem: one call, one ordering. row_limit goes up to 25000. Use the cursor for browsing, or when the result genuinely exceeds one page, and treat a crawled set as approximate rather than exact. Note: Data freshness lag — recent dates (last 3+ days) may be incomplete. Check firstIncompleteDate in response.

Get top queries with clicks, impressions, CTR, and position. Automatically identifies high-opportunity queries (position 4-20, impressions >100, CTR <5%). Args: site_url (string, required): The Search Console property URL. days (int, optional): Number of days to look back. Range: [1, 90]. Default: 28. start_date / end_date (string, optional): Explicit date range (overrides days). page_filter (string, optional): Filter to queries for a specific URL (exact match). device (string, optional): Filter by device type. Valid: desktop, mobile, tablet. search_type (string, optional): web (default), image, video, news. row_limit (int, optional): Queries per page. Default: 25. cursor (string, optional): Opaque cursor from a previous response's pagination.nextCursor. Omit for the first page; never construct one by hand. Returns: data: top query rows for this page highOpportunityQueries: queries with ranking potential (position 4-20, impressions >100, CTR <5%), derived from this page only — not from the whole result set rowCount: rows in this page pagination: { hasMore, nextCursor? } — nextCursor is absent on the last page site_url, dateRange Pagination: when pagination.hasMore is true, call again with the same parameters plus cursor set to pagination.nextCursor. A cursor only continues the exact call that returned it, including the date window it resolved to — so a crawl that spans midnight on a relative range (days) is rejected rather than paging a window that has since shifted. Start again from the first page if that happens. IMPORTANT — when you need a complete, consistent result set, raise row_limit and take it in ONE call rather than crawling. Search Console sorts by clicks descending and orders ties "in an arbitrary way", and that tie order is not stable between calls; on a typical property most queries have 0 clicks and are therefore all tied, so a multi-page crawl can repeat some and miss others. Use the cursor for browsing, and treat a crawled set as approximate rather than exact.

Compare current vs previous period performance in a single call. Makes 4 parallel API calls internally and returns deltas for clicks, impressions, CTR, and position. Args: site_url (string, required): The Search Console property URL. days (int, optional): Length of each period in days. Range: [1, 90]. Default: 28. current_end_date (string, optional): Override the end date of the current period (YYYY-MM-DD). Default: today minus 3 days. search_type (string, optional): web (default), image, video, news. Returns: currentPeriod, previousPeriod: { startDate, endDate, clicks, impressions, ctr, position } delta: { clicks, clicksPct, impressions, impressionsPct, ctr, position } trend: "improving" (clicks up >10%), "declining" (clicks down >10%), or "stable" topQueries, previousTopQueries: top 50 queries for each period (may be null if sub-query failed) errors: list of sub-query failures (if any) — overall metrics are always present Periods are adjacent and non-overlapping: previousEnd + 1 day = currentStart.

Get a complete overview of search performance in a single call: overall metrics, top queries, top pages, and daily trend. Makes 4 parallel API calls internally. Args: site_url (string, required): The Search Console property URL. days (int, optional): Number of days to look back. Range: [1, 90]. Default: 28. search_type (string, optional): web (default), image, video, news. Returns: overall: aggregate metrics { startDate, endDate, clicks, impressions, ctr, position } topQueries: top 10 queries by clicks topPages: top 10 pages by clicks byDate: daily breakdown dateRange: { startDate, endDate } errors: list of sub-query failures (if any) — partial results are returned when sub-queries fail

Troubleshoot authentication issues. Only use this when another tool fails with a permission or auth error — do NOT call proactively. Returns: authenticated status, user_id, email, and a human-readable message.

Why a call was refused with a billing error (402) and how to get unblocked. Reports whether this organization is entitled, its subscription status, tier, trial end, credits used and left this period, the price table and a link to the billing page. With checkout_tier, the organization owner also gets a Stripe Checkout link for that plan; a member gets the owner's contact and a message to forward. Call it after any 402. While waiting for a human to pay, poll it every 10 seconds for the first minute, every 30 seconds until 5 minutes, then every 60 seconds until 15 minutes, then stop and tell the human. Once it returns entitled: true, retry the blocked call; if the retry still returns 402, wait 15 seconds and retry once. A Checkout link is a payment link: give it to the organization owner only.

List the Google Search Console connections available to you — both ones you own and ones shared with you via an organization. Only Google connections with Search Console (webmasters.readonly) scope are returned. Use this to discover connection IDs for set_default / share / rename / revoke.

Set the Google Search Console connection that should be used by default for subsequent Google Search Console tool calls. The default is scoped to the calling actor (your user account, or the API key being used).

Share an owned Google Search Console connection with all members of your organization, so teammates can use it without having to reconnect Google themselves. You must be the owner of the connection.

Stop sharing an owned Google Search Console connection with your organization. Teammates lose access immediately. You must be the owner.

Rename the display name of an owned Google Search Console connection. The OAuth grant and underlying account are unaffected — this only changes the human-readable label. You must be the owner.

Revoke (soft-delete) an owned Google Search Console connection. Any defaults pointing to it are invalidated and shared org members lose access. The OAuth grant at Google is NOT revoked by this tool — the user must disconnect via the dashboard if they want to fully revoke at Google. You must be the owner.

Submit feedback about missing tools, improvements, bugs, or workflow gaps in the Search Console MCP toolset. Not for user-facing issues (auth errors, API errors). Args: feedback_type: new_tool, improvement, bug, workflow_gap title: Short description (5-200 chars) description: What is needed and why (20-2000 chars) current_workaround (optional): How you're working around the gap priority (optional): low, medium (default), high Returns: confirmation that feedback was recorded.

Health check. Returns server status, version, and timestamp.

List all Google Search Console properties (sites) the user has access to. Use this tool first to discover which properties are available before calling other tools. Results are cached for 15 minutes; pass refresh=true only when you need the latest list. Args: refresh (bool, optional): Force a fresh fetch from the API, bypassing cache. Default: false. Returns: data: every site property, with siteUrl and permissionLevel (siteOwner, siteFullUser, siteRestrictedUser, siteUnverifiedUser). cached, synced_at Not paginated: the Search Console sites endpoint takes no page size or page token and returns the whole list at once, so this single call is already the complete set — there is no cursor to follow.

List sitemaps submitted to Google Search Console for a property. Args: site_url (string, required): The Search Console property URL (e.g. https://example.com/ or sc-domain:example.com). refresh (bool, optional): Force a fresh fetch from the API, bypassing cache. Default: false. Returns: data: every submitted sitemap, with path, last submission date, pending/error status, and error/warning counts. site_url, cached, synced_at Not paginated: the Search Console sitemaps endpoint takes no page size or page token and returns the whole list at once, so this single call is already the complete set — there is no cursor to follow.

Get detailed indexing and crawl status for a specific URL using the Google Search Console URL Inspection API. Args: inspection_url (string, required): Fully-qualified URL to inspect. Must be under the site_url property. site_url (string, required): The Search Console property that contains the URL (e.g. https://example.com/ or sc-domain:example.com). language_code (string, optional): IETF BCP-47 language code for translated issue messages (e.g. en-US). Returns: verdict (PASS/FAIL/NEUTRAL), coverage state, robots.txt state, indexing state, page fetch state, last crawl time, crawled as (mobile/desktop), Google canonical URL, user canonical URL, sitemap references. Note: Use google_search_console_list_sites first to discover valid site_url values.