Optimly MCP

MCP tool reference

Product

Optimly MCP

Updated

2026-09-14

• brand_action_plan — Generate prioritized recommendations for improving how AI represents a brand, grounded in its registry profile and source-graph findings.
• brand_compare — Compare a primary brand with up to five competitors on AI representation, archetype, Brand Authority Index tier, sentiment, positioning, and key discrepancies.
• brand_lookup — Look up a brand in the Optimly AI Brand Index by name or slug. Returns its public registry identity, Brand Authority Index tier, archetype, category, and profile URL.
• brand_monitor — Report score changes, narrative drift, competitor movement, and alerts since a selected date.
• brand_narrative_check — Return the current AI-perception narrative for a brand, including positioning, sentiment, accuracy gaps, competitive associations, claims, and sources when available.
• brand_publication_expand — Find lookalike publications the brand does not yet win, labeling candidates by discovery source and verification level.
• brand_publication_match — Rank third-party publications and the category angles they own in the brand's citation graph, including authority, model breadth, accuracy, and pitch signals.
• brand_publish — Execute one approved brand_vault.publish action that atomically updates any combination of BrandVault text sections, sub-brands, structured competitors, and disambiguation. The publish also refreshes the brand's canonical Optimly-hosted llms.txt.
• brand_site_audit — Audit key website pages from an AI crawler's perspective: rendering, AI-bot robots rules, structured data, metadata, llms.txt, status, and visible content.
• brand_sources — Aggregate citations from an audit and classify sources as owned or third-party, with grounded-answer rates and accuracy comparisons.
• brand_status — Return Business Profile baseline version, last-edit timestamps, and section summaries for the authorized brand.
• brand_update — Execute one exactly approved Business Profile mutation. The call atomically starts the action and consumes its approval.
• brand_vault_profile — Read all 15 verified ground-truth sections for the authorized brand.
• check_action_idempotency — Check whether an operation and idempotency-key pair was claimed and whether a durable response exists.
• competitive_analysis — Return the authorized brand's latest private competitive analysis across AI models.
• coverage_add — Add one earned-media article URL to your brand's Media Coverage intake. Optimly normalizes the URL, enriches missing article details, and saves it for citation matching. This is a distinct ingestion flow, not a BrandVault section update.
• coverage_add_many — Bulk-add 1 to 25 earned-media article URLs to your brand's Media Coverage intake. Optimly enriches missing details, deduplicates normalized URLs, and returns added, already_tracked, or failed for every item.
• coverage_list — List the earned-media URLs currently tracked for your brand, including publication details and whether each URL has appeared as a citation in monitored AI answers.
• create_brand_vault_publish_action — Create one immutable approval proposal that can update every editable BrandVault surface: text sections, sub-brands, structured competitors, and disambiguation. Supplied structured collections are full replacements; omitted collections are preserved.
• create_execution_action — Create an immutable, payload-hashed action proposal before requesting approval or recording execution. For a single Business Profile field use capability brand_vault_update with operation brand_vault.update_section and exactly { section_key, content }.
• delta_report — Return the latest or a selected Brand Authority Index audit summary, including run metadata and component scores.
• delta_report_details — Return per-question, per-model answers from a completed audit, including answer text, expected ground truth, accuracy, and owned versus third-party citations.
• delta_report_history — List completed audit runs newest-first with component scores and run-over-run deltas.
• get_action_status — Read an execution action and its append-only status history.
• get_actions — Read recent governed actions and their independent verification state without approving or executing work.
• get_agent_activity — Read tenant-scoped AI agent and crawler request activity for a connected site. Returns bounded totals, agent families, verified counts, requested paths, and observation state. Declared identities are not verified.
• get_agent_content_gaps — Read ranked paths that AI agents or crawlers tried to retrieve but could not successfully access, with observed outcomes, operators, purpose, identity confidence, capture fidelity, timestamps, closest existing paths, and representative evidence.
• get_agent_work — Read the selected brand's current recommendations, drafts, queue placement, dismissals, and checklist state before proposing new work. This tool never changes work.
• get_ai_representation — Read current AI Representation KPIs, archetype, and bounded question and model evidence.
• get_attribution_evidence — Read actions, AI Representation movement, and website outcomes together as directional attribution evidence. Never treats sequence or correlation as causal proof.
• get_site_readiness — Read current discovery-readiness checks for the connected website, including crawler access, robots, sitemap, canonical, schema, and llms.txt.
• get_site_retrievals — Read grounded model retrieval and citation evidence for owned website pages.
• get_whitespace_candidates — Return ranked information gaps for the selected brand, with observed demand, supply weakness, corroboration, score, status, and raw supporting evidence.
• get_work_status — Read one agent work item, its producer provenance, evidence, revisions, checklist, and feedback.
• inspect_agent_view — Inspect the latest observed Agent Lens response evidence for one normalized public path. Returns operator, purpose, identity confidence, status, outcome, redirects, capture method and fidelity, and a snapshot reference when available. It does not return page content.
• inspect_website — Read the selected brand's connected website snapshot before making website, traffic, conversion, referral, crawler, schema, or site-analysis recommendations. Returns installation health, bounded observed metrics, data-quality warnings, and evidence references.
• list_accessible_brands — List every brand the current Optimly OAuth user may access, including organization and permission scope.
• list_my_brands — List the brands the current Optimly OAuth user may access. Use this before a brand tool when more than one brand is available.
• list_pending_approvals — List pending approval requests for the selected brand.
• magic_moment_analysis — Deliver the authenticated brand's one-time Optimly value analysis. Combines the Business Profile, latest AI Representation evidence, and connected website outcomes, then identifies exactly one highest-value gap and one concrete next action.
• record_action_failed — Record that an external execution action failed. BrandVault completion is recorded by brand_update or brand_publish and does not require this tool.
• record_action_started — Record that an external execution action started. Never use for BrandVault updates: brand_update and brand_publish perform the start and approval consumption atomically.
• record_action_succeeded — Record that an external execution action succeeded. BrandVault completion is recorded by brand_update or brand_publish and does not require this tool.
• record_agent_feedback — Record explicit customer feedback on a recommendation or draft. Use the customer's stated reason and explanation; never invent feedback or treat silence as dismissal.
• record_approval_decision — Attest an approver decision with external evidence.
• record_checklist_progress — Record explicit checklist progress reported by the customer or independently evidenced. A human completion statement is not independent verification; include public evidence only when it exists.
• register_brand_approver — Register or update the single verified approval contact for the selected brand.
• request_action_approval — Request approval for one or more exact proposed action payloads from the registered brand approver.
• revoke_action_approval — Revoke a pending or unused approval before execution starts.
• save_draft — Save one copy-pasteable, evidence-backed draft to the same Optimly Actions workflow. The draft may contain portable website copy, profile copy, structured data, or an outreach pitch, but must not contain invented facts or credentials.
• save_recommendation — Save one evidence-backed onsite or offsite recommendation to the shared Optimly Actions inbox.
• simulate_agent_preview — Run a controlled, read-only fetch of one configured public path using a declared AI agent identity. Returns status, outcome, redirects, bounded representation diagnostics, and a content fingerprint. This is simulated evidence, not observed live traffic or proof of a private agent's behavior.
• update_candidate_status — Pin, target, hold, dismiss, or reopen a whitespace candidate. Targeted candidates require a durable action reference for attribution.