Use MCP with full access

Use up to 13 scoped tools across an eligible paid account for Search Console, linked GA4, SEO planning data, and human-approved manual annotation changes.

Product guide · Updated August 2026

Understand full MCP access

Eligible paid access lets an owner or member approve up to seven scopes for the selected CrawlRaven account. A paid grant is not pinned to one free-preview site and is not subject to the free preview’s 20-call daily allowance. It remains bound to one client, user, and account. Most tools are read-only; manual annotation changes require the separate Manage annotations scope and same-user approval for every proposal.

ScopeWhat it provides
websites:readWebsite IDs, names, lifecycle state, activation, and full-workspace access.
search_performance:readOverview, daily totals, and bounded Query and Page results.
opportunities:readPersisted opportunities, Priority, evidence, and recommended actions.
targets:readDurable target plans, intended URLs, groups, status, and imported research fields.
annotations:readActive manual notes, Google updates, and opportunity milestones from the timeline.
analytics:readBounded reports from the GA4 property already linked to the website.
annotations:writePlan manual note create, update, archive, or restore operations for same-user approval, check plan status, and list archived notes.

The complete paid catalog contains the eight core tools plus get_analytics_summary, get_analytics_breakdown, plan_annotation_change, get_annotation_change_status, and list_archived_annotations. A client sees only the subset covered by its approved scopes.

Connect with https://mcp.crawlraven.com/mcp, complete browser OAuth, review the selected account and scopes, then begin with list_websites. Use the returned website_id for every site-specific tool. Existing connections are never widened silently; reconnect to add View analytics or Manage annotations.

Build workflows that chain tools

Replace [your site] with the exact CrawlRaven site name or domain in the full-access prompts below.

WorkflowPrompt
Monthly search reviewFor [your site], compare the last 90 days with the preceding period, identify the returned pages and queries with the largest click losses, and check the timeline for nearby notes or Google updates.
Decay triageFind returned pages that lost clicks over 90 days, match them to current content-decay opportunities, and rank the resulting actions by Priority and current page traffic.
Page refresh researchFor [your site], list the queries attributed to exactly [page URL] from [start date] through [end date]. Combine them with the page's current public content, then propose refresh ideas grounded in the returned query evidence. Keep query evidence separate from content observations and state any truncation.
Planning coverageCompare targeted keywords with the bounded top-query corpus, then flag targets with no planned URL or no matching returned query. State the evidence boundary.
Search-to-outcome reviewCombine the last 28 days of Search Console movement with linked GA4 landing-page sessions and key events for the same complete-day period. Keep query evidence separate from page-level outcomes and report truncation.
Record a reviewed changeAfter reviewing the evidence, plan a concise manual timeline note for [date]. Show the exact values and approval link, and do not say it changed until I approve it in CrawlRaven and status confirms completion.
Weekly backlogTurn the highest-priority active opportunities into a five-item implementation brief with the reason, evidence, affected page, and recommended action.

Browse more copyable examples for founders, SEO specialists, agencies, content teams, and developers in Prompts by role.

Choose windows and page only when needed

Tool groupDefaultMaximum and window
Query and Page lists25 rows100 rows per page; rolling 1–365 days or exact inclusive dates spanning at most 365 days.
Opportunities and Target Keywords25 rows100 rows per page; no reporting-window parameter.
Active annotations100 rows200 rows per page; choose 1 through 400 days.
Archived annotations25 rows100 rows per page; cursor-paginated and not time-bound.
Analytics summary28 complete daysRolling or exact windows up to 365 complete UTC days; one result with up to 365 daily points.
Analytics breakdown25 top rowsUp to 100 rows; no cursor, so inspect truncated and narrow the dimension or filters.
Annotation plan or statusOne resultPlans expire after 15 minutes; status lookup does not repeat a mutation.
WebsitesAll returnedUp to 100 websites; no cursor pagination.
Overview and daily search performance28 daysRolling 1–365 days or exact inclusive dates spanning at most 365 days.

When next_cursor is present, pass it back unchanged as cursor with the same site, window, search, exact page filter, and status. Changing that scope invalidates the cursor. Ask the agent to stop when the cursor is empty and to report any truncation rather than implying the source was exhaustive.

Know what full access still cannot do

  • It cannot change Search Console, Analytics properties, sites, targets, opportunity states, billing, or administration; run exports; or call arbitrary API routes.
  • It cannot refresh Opportunities or fetch arbitrary URLs, files, webhooks, or email.
  • Analytics tools read only the GA4 property already linked in CrawlRaven. They cannot connect or switch a property, expose credentials, or change Google configuration.
  • Manage annotations can only propose create, update, archive, or restore for manual notes. The same authorizing user must approve every exact plan in CrawlRaven within 15 minutes.
  • Update, archive, and restore are version-checked; stale plans conflict. Create retries are idempotent within the documented 90-day retention window.
  • Derived and imported events are immutable. MCP can archive and restore manual notes but can never permanently delete one.
  • Target Keyword results contain the durable plan, not live observed-page or mismatch evidence. Use the app for live target evidence.
  • Paid list_queries accepts an exact absolute HTTP(S) page URL and can combine it with query-text search. This extends the existing tool; it is not a separate page-query tool. Query and Page tools still do not expose the app’s country, device, or branded filters.

Tool results leave CrawlRaven for the connected client and model provider. Review their data policies, connect only clients you trust with Search Console, linked GA4, and CrawlRaven planning data, and revoke old connections from MCP connections. Grants expire after 90 days. If paid access lapses, the grant remains visible and revocable but returns no data until paid access is restored.

See the MCP server reference for exact parameters, client commands, OAuth behavior, and error recovery.