generated: '2026-08-13' method: derived source: https://developers.scrunch.com/mcp/tools note: > Binds each of the 33 published Scrunch MCP ("Scrunchie") tools to the backing Scrunch Data API operationId(s). The live MCP tools/list is OAuth-gated (POST https://app.scrunchai.com/api/mcp returns 401 invalid_token anonymously), so no inputSchema was retrievable from the server. Tool names and descriptions are taken verbatim from the provider's published tool reference; the REST bindings are mapped by name and documented semantics against operationIds verified to exist in openapi/_original/scrunch-ai-data-api-runtime-openapi.json and openapi/_original/scrunch-ai-data-api-openapi.yml. Confidence is set honestly - nothing here was fabricated, and the four metrics tools bind to the single query operation because that is what the docs describe them as projecting. surfaces: openapi: files: - openapi/_original/scrunch-ai-data-api-runtime-openapi.json - openapi/_original/scrunch-ai-data-api-openapi.yml base_url: https://api.scrunchai.com/v1 gated: false note: > Two non-identical projections. The live runtime spec (37 ops) carries signals, getBrand, archiveBrand, getPrompt and createCollection; the docs-derived spec (33 ops) carries the orchestration and ai-referrals operations the runtime spec omits. The union is 44 operationIds. graphql: present: false mcp: url: https://app.scrunchai.com/api/mcp transport: streamable-http auth: oauth gated: true probe: method: tools/list status: 401 body: invalid_token checked: '2026-08-13' coverage: tools_published: 33 tools_bound_to_rest: 18 mcp_only: 15 rest_operations_total: 44 rest_operations_with_a_tool: 18 crosswalk: - tool: get_presence_metrics category: reporting rest: [query] binding: rest confidence: medium note: Brand mention rate over time - a metric projection of the single aggregate query operation. - tool: get_position_metrics category: reporting rest: [query] binding: rest confidence: medium note: Top/middle/bottom position share - same query operation, different metric selection. - tool: get_sentiment_metrics category: reporting rest: [query] binding: rest confidence: medium note: Positive/mixed/negative breakdown - same query operation. - tool: get_citations_metrics category: reporting rest: [query] binding: rest confidence: medium note: Citation share by owner (brand/competitor/third-party) - same query operation. - tool: list_responses category: responses rest: [listResponses] binding: rest confidence: high note: MCP defaults to ids_only with limit 10 / max 100; the REST operation returns full records. - tool: list_prompt_variants category: prompts rest: [listPrompts] binding: rest confidence: high - tool: create_prompt category: prompts rest: [createPrompt] binding: rest confidence: high - tool: archive_prompt category: prompts rest: [archivePrompt] binding: rest confidence: high - tool: list_brands category: configure rest: [listBrands] binding: rest confidence: high - tool: create_brand category: configure rest: [createBrand] binding: rest confidence: high - tool: update_brand category: configure rest: [patchBrand] binding: rest confidence: high - tool: list_competitors category: configure rest: [listCompetitors] binding: rest confidence: high - tool: create_competitor category: configure rest: [createCompetitor] binding: rest confidence: high - tool: update_competitor category: configure rest: [updateCompetitor] binding: rest confidence: high - tool: archive_competitor category: configure rest: [archiveCompetitor] binding: rest confidence: high - tool: list_personas category: configure rest: [listPersonas] binding: rest confidence: high - tool: create_persona category: configure rest: [createPersona] binding: rest confidence: high - tool: archive_persona category: configure rest: [archivePersona] binding: rest confidence: high - tool: export_agent_traffic category: agent-traffic rest: [getAgentTraffic] binding: rest confidence: high note: Exports the aggregated bot/agent traffic the REST operation returns (CSV-shaped for spreadsheets). mcp_only: - tool: generate_explorer_chart reason: > Builds a Scrunch Explorer chart spec and returns a deep link into the app. No public REST operation creates Explorer charts. - tool: create_explorer_dashboard reason: > Saves an Explorer dashboard of metric tiles and returns a deep link. The runtime spec's createCollection is not documented as the dashboard endpoint, so no binding is asserted. - tool: get_response reason: > Fetches a single AI response by ID. The REST surface publishes listResponses only - there is no getResponse operation in either spec. - tool: unarchive_prompt reason: REST publishes archivePrompt but no unarchive operation. - tool: get_tags reason: No REST operation exposes the brand tag list; tags appear only as fields on prompts. - tool: add_tag_to_prompts reason: Bulk tag append across up to 500 prompts; no public REST operation. - tool: remove_tag_from_prompts reason: Bulk tag removal; no public REST operation. - tool: update_persona_on_prompts reason: > Reassigns a persona across prompts. updatePersona edits the persona record itself, not its assignment across prompts, so no binding is asserted. - tool: import_agent_traffic reason: > Ingests agent traffic. The documented ingestion endpoint is POST https://webhooks.scrunchai.com/v1/sites/{site_id}/platforms/custom/web-traffic, which sits on a different host and appears in no OpenAPI document, so it has no operationId to bind to. - tool: list_scrunch_help_collections reason: Scrunch help-center content; not part of the Data API. - tool: get_scrunch_help_article reason: Scrunch help-center content; not part of the Data API. - tool: fetch_scrunch_url reason: Server-side fetch of a Scrunch app/deep-link URL; no REST equivalent. - tool: get_caller_identity reason: > Returns the authenticated MCP caller's identity. The REST surface has only root__get, which is an unlabelled root probe, not an identity endpoint. - tool: get_current_date reason: Server-side clock utility for relative date phrasing; not an API capability. rest_only: - capability: Signals (nightly detection sweep) operations: [listSignals, getSignal, listSignalAnchors, listSignalReactions, setSignalReaction, clearSignalReaction] note: > Newest REST surface, live in the runtime spec and documented at /api-reference/signals/overview. No MCP tool exposes it as of the published 33-tool reference. Reaction writes require a user JWT - API keys receive 403. - capability: Optimize-and-Deploy orchestration operations: [createOptimizeAndDeployPipeline, getOptimizeAndDeployStatus, bulkRenderOptimizedPages] note: The AXP pipeline is REST + webhooks only; no MCP tool drives it. - capability: Site Audit operations: [createPageAudits, listPageAudits, getPageAudit] - capability: Sitemap operations: [listSitemapPages, getSitemapPage, getSitemapPageMetrics, exportSitemapPages] - capability: AI Referrals ingestion operations: [registerAiReferralsConnection, listAiReferralsConnections, pushAiReferralsEvents, disableAiReferralsConnection] - capability: Read-side configuration operations: [getBrand, archiveBrand, getPrompt, getCompetitor, getPersona, updatePersona, createCollection] - capability: Service root operations: [root__get]