add_action_plan_step
ChatGPTAdd one custom step to the workspace action plan. Requires a title and a description of why. Does not edit the shared playbook catalog.
add_prompt
ChatGPTTrack a new question. Over the plan active cap, or without a paid plan, the prompt is saved inactive and added_as_inactive is true. Text is 2 to 200 characters. Type is commercial, branded, or competitor.
analyze_action_plan
ChatGPTWhich of this month’s GEO checklist steps are still open and why they are in focus.
analyze_ai_recommendation_score
ChatGPTCompare the workspace AI recommendation score with the previous point and name the weakest of the eight components. No extra LLM.
analyze_prompt
ChatGPTDeterministic briefing for one tracked question: mention/citation rates, top hosts, latest answers.
analyze_site
ChatGPTFetch the public website of the workspace domain and run a bounded live audit (robots, llms.txt, sitemap sample, homepage JSON-LD/SEO). Not a full-site crawl.
analyze_sources
ChatGPTCited vs consulted domains and citation gaps (PR vs community-style actions).
analyze_weekly_report
ChatGPTOne markdown briefing: weekly visibility, rivals, and open action-plan steps.
check_ai_crawler_robots
ChatGPTFetch the public website for one domain and run only this check: robots.txt allows the tracked AI crawlers. A pass means robots.txt allows the tracked AI crawlers. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_ai_txt_and_llms_full
ChatGPTFetch the public website for one domain and run only this check: ai.txt and llms-full.txt exist and are not disallowed. A pass means ai.txt and llms-full.txt exist and are not disallowed. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_canonical_and_hreflang
ChatGPTFetch the public website for one domain and run only this check: a canonical link and hreflang are present. A pass means a canonical link and hreflang are present. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_citation_signals
ChatGPTFetch the public website for one domain and run only this check: date, author, and publisher signals are present. A pass means date, author, and publisher signals are present. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_homepage_not_noindex
ChatGPTFetch the public website for one domain and run only this check: the homepage is not marked noindex. A pass means the homepage is not marked noindex. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_homepage_visible_text
ChatGPTFetch the public website for one domain and run only this check: the homepage has visible text without a hydration gap. A pass means the homepage has visible text without a hydration gap. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_json_ld
ChatGPTFetch the public website for one domain and run only this check: structured data includes an entity type such as Organization or FAQPage. A pass means structured data includes an entity type such as Organization or FAQPage. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_llms_txt
ChatGPTFetch the public website for one domain and run only this check: /llms.txt exists, is long enough, and is not disallowed. A pass means /llms.txt exists, is long enough, and is not disallowed. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_llms_txt_links
ChatGPTFetch the public website for one domain and run only this check: links in llms.txt on the same site respond. A pass means links in llms.txt on the same site respond. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_open_graph_tags
ChatGPTFetch the public website for one domain and run only this check: Open Graph title, description, and image are present. A pass means Open Graph title, description, and image are present. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_question_headings
ChatGPTFetch the public website for one domain and run only this check: question headings and FAQPage are present. A pass means question headings and FAQPage are present. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_robots_tag_agreement
ChatGPTFetch the public website for one domain and run only this check: meta robots and X-Robots-Tag agree on noindex. A pass means meta robots and X-Robots-Tag agree on noindex. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_single_h1
ChatGPTFetch the public website for one domain and run only this check: the page has exactly one H1. A pass means the page has exactly one H1. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_sitemap_freshness
ChatGPTFetch the public website for one domain and run only this check: the newest sitemap lastmod is within 90 days. A pass means the newest sitemap lastmod is within 90 days. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_sitemap_urls
ChatGPTFetch the public website for one domain and run only this check: a sitemap exists and its URLs respond. A pass means a sitemap exists and its URLs respond. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
check_title_and_description
ChatGPTFetch the public website for one domain and run only this check: title and description sit in the target length bands. A pass means title and description sit in the target length bands. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
compare_competitors
ChatGPTRank competitor hosts by mentions and citations from the current visibility snapshot, and contrast them with the brand.
complete_action_plan_step
ChatGPTMark one existing action-plan step complete or incomplete. Requires the step id, the completed flag, and a description of why. Does not edit the shared playbook catalog.
connect_blog_domain
ChatGPTReturn the CNAME the customer adds at their domain host so blog.{domain} serves their content posts. Does not change DNS and does not change the store address. Walk the user through the returned steps. The name is blog, without the rest of the domain. The target is the same for every customer.
create_post_with_agent
ChatGPTStart the content agent for one content Plan of Action step. Spends 2 credits and fills the next open daily post time. Pass preferred_day only when the user chose a calendar day. Take plan_step_id from get_action_plan. The step must already be marked as content. Does not publish by itself.
edit_action_plan_step
ChatGPTOverwrite the description of one existing action-plan step. Requires the step id and a description of why. Does not edit the shared playbook catalog.
enqueue_visibility_scan
ChatGPTEnqueue an AI search mention scan for a workspace. Consumes plan quota. Pass an Idempotency-Key or let the server generate one. Prefer wait_for_ai_search_scan to poll automatically.
explain_visibility_gaps
ChatGPTList tracked prompts where the brand has no mention but at least one competitor does in the current Project visibility snapshot.
export_blog_content
ChatGPTDownload the workspace posts as WordPress WXR, Blogger Atom, or RSS. The file is the same shape the importer reads. Does not change the blog.
get_action_plan
ChatGPTThis month’s GEO checklist (up to four focus steps), custom steps, and completion status.
get_ai_recommendation_score
ChatGPTRead the AI recommendation score for a workspace: the headline score, eight components, and the time series. Use analyze_ai_recommendation_score to compare it with the previous point.
get_ai_sources
ChatGPTCited and consulted domains, fan-out queries, and citation gaps. Prefer analyze_sources.
get_blog_health
ChatGPTRead the customer blog health checks: domain, SSL, last publish, access, posts, keywords, SEO, GEO, AIO, AEO, schema, theme, robots.txt, sitemap.xml, and llms.txt.
get_blog_settings
ChatGPTRead the customer blog template, theme, font, uploaded font URL, colors, logo, author, social links, pixels, three CTAs, footer switch, categories, tags, daily post times, and publish policy. layout is the template. preset is the theme. Does not change the blog.
get_blog_telemetry
ChatGPTRead published post telemetry: impressions, direct, SEO, AI, CTA clicks, total, CTR, and which AI products visited.
get_brand_perception
ChatGPTAttributes AI answers associate with the brand vs competitors (0–100).
get_content_calendar
ChatGPTRead scheduled and published content posts for the workspace: day, time, state, status, title, and the daily post times. Does not change a slot.
get_content_job
ChatGPTRead one content job: state, subject, slug, scheduled day and time, and the draft when it is ready for review. Does not publish.
get_crawl_insights
ChatGPTAI crawler hit aggregates by bot and path, plus how often those paths were cited. Not raw log files.
get_doc
ChatGPTReturn the full markdown for a public MencionAI guide. Use list_docs for ids (api, sdk, whitelabel, ars, mcp, mcp-cursor, mcp-claude, mcp-chatgpt, mcp-codex, mcp-antigravity, mcp-n8n, llms). Bundled guides ship with the package; llms is fetched live from mencionai.com. Does not require an API key.
get_job
ChatGPTGet status for a visibility job previously enqueued via enqueue_visibility_scan.
get_onboarding_v2_report
ChatGPTRead the AI search report for one onboarding prompt batch: where answers mention the brand, which competitors appear, and which sources are cited.
get_organization
ChatGPTReturn the authenticated MencionAI organization (id, name, plan, subscription status, covered_providers).
get_project_settings
ChatGPTBrand name, domain, and competitor list. Does not include site keys, API secrets, or billing.
get_prompt
ChatGPTPrompt detail: visibility, top cited hosts, and latest answers. Answer snippets require visibility:answers:read (stripped to null without it).
get_traffic_series
ChatGPTDaily AI referral visit counts for a workspace (range 7d, 30d, or 90d; default 30d).
get_traffic_summary
ChatGPTObserved visits from ChatGPT, Gemini, and other AI assistants on the workspace domain (requires the site snippet). Range 7d/30d/90d.
get_visibility_citations
ChatGPTPower-user fallback — prefer get_visibility_overview for a short citation sample. Domains and URLs from the current Project visibility snapshot. Omit limit to return every source.
get_visibility_competitors
ChatGPTPower-user fallback — prefer compare_competitors for ranked brand vs rivals. Competitor hosts and mention/citation counts.
get_visibility_coverage
ChatGPTPower-user fallback — prefer get_visibility_overview or explain_visibility_gaps. Per-prompt mention coverage (mentioned/cited flags and counts).
get_visibility_mentions
ChatGPTPower-user fallback — prefer get_visibility_overview. Brand mentions from the current visibility snapshot, grouped by tracked prompt text.
get_visibility_overview
ChatGPTStart here for agent workflows: Project visibility snapshot with summary, mention coverage, competitors, and citation sources for a workspace.
get_visibility_responses
ChatGPTAI answer excerpts from the visibility snapshot. Requires API key scope visibility:answers:read. Fails closed if the scope is missing.
get_visibility_series
ChatGPTDaily AI/SEO/visibility score series for a workspace (range 7d, 30d, or 90d; default 30d).
get_visibility_summary
ChatGPTPower-user fallback — prefer get_visibility_overview. Current Project visibility snapshot summary (prompts, mentions, citations).
get_weekly_report
ChatGPTLive weekly report summary: visibility, mention coverage, competitor stats. Prefer analyze_weekly_report for a briefing.
get_workspace
ChatGPTGet a single workspace by id (brand name, domain).
identify_site_stack
ChatGPTFetch the public website of the workspace domain, detect its stack (WordPress, Next.js, Vercel, Lovable, Replit, Shopify, Webflow, Wix, and similar) from live headers and HTML, and overwrite the saved stack on domain settings.
import_blog_content
ChatGPTImport a WordPress WXR, Blogger Atom, or RSS/Atom file into the workspace as draft posts. xml is the file text, at most 4 MB. The same source id is skipped on a later import. Does not publish and does not spend credits.
ingest_crawl_insights_csv
ChatGPTStore a user-provided access-log CSV (timestamp, user_agent, path, status). Requires workspaces:write. Do not invent rows.
list_content_hosts
ChatGPTFind the apex, www, blog subdomain, MencionAI platform blog, homepage-linked hosts, and content labels such as docs, content, and help when those hosts publish a sitemap. Returns each host and a sample of its sitemap URLs. A host in the list is a site to check. Run site_check_* on that host next.
list_docs
ChatGPTList MencionAI public documentation guides available via get_doc (API, SDK, whitelabel, ARS, MCP, Cursor/Claude/ChatGPT/Codex/Antigravity/n8n, llms.txt). Use for how-to questions before inventing answers. Does not require an API key.
list_prompts
ChatGPTList tracked questions (prompts) for a workspace: text, type, active flag, position, volume.
list_site_pages
ChatGPTFetch the public sitemap of the workspace domain and list every URL (blog, content, product, other), including a blog with 100 pages. Also fetches titles, H1, JSON-LD, and link counts until the time budget ends. Not a backlink index. Use analyze_sources for AI citations.
list_workspaces
ChatGPTList workspaces (brand/domain projects) for the organization. Omit limit to return every workspace. Pass limit and follow next_cursor to walk a long organization.
prepare_geo_content_brief
ChatGPTAssemble a writing brief for a blog or social post using visibility gaps, per-model search-backed directions, the GEO action plan, and locale. For blog, include the citation contract (GEO/AEO/SEO/AIO). Call research_geo_practices first, web-search, then validate_geo_content after drafting.
reschedule_content_job
ChatGPTMove an unpublished content draft to another calendar day and clock time. A past day or a taken slot is rejected. Does not publish.
research_geo_practices
ChatGPTBuild a host-LLM research task with live-search queries and official URLs for every MencionAI-tracked engine (ChatGPT, Gemini, Google AI Studio, Claude, Grok, DeepSeek, Perplexity). Customer drafts should still name only plan-covered models. You must web-search. Do not invent crawler tokens. Call this before prepare_geo_content_brief.
retry_content_job
ChatGPTRewind a failed or blocked content job to the step that failed and run it again. Does not debit credits again. A job that is not failed or blocked is rejected.
site_check_ai_crawler_robots
ChatGPTFetch the public website for one domain and run only this check: robots.txt allows the tracked AI crawlers. A pass means robots.txt allows the tracked AI crawlers. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_ai_txt_and_llms_full
ChatGPTFetch the public website for one domain and run only this check: ai.txt and llms-full.txt exist and are not disallowed. A pass means ai.txt and llms-full.txt exist and are not disallowed. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_canonical_and_hreflang
ChatGPTFetch the public website for one domain and run only this check: a canonical link and hreflang are present. A pass means a canonical link and hreflang are present. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_citation_signals
ChatGPTFetch the public website for one domain and run only this check: date, author, and publisher signals are present. A pass means date, author, and publisher signals are present. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_homepage_not_noindex
ChatGPTFetch the public website for one domain and run only this check: the homepage is not marked noindex. A pass means the homepage is not marked noindex. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_homepage_visible_text
ChatGPTFetch the public website for one domain and run only this check: the homepage has visible text without a hydration gap. A pass means the homepage has visible text without a hydration gap. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_json_ld
ChatGPTFetch the public website for one domain and run only this check: structured data includes an entity type such as Organization or FAQPage. A pass means structured data includes an entity type such as Organization or FAQPage. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_llms_txt
ChatGPTFetch the public website for one domain and run only this check: /llms.txt exists, is long enough, and is not disallowed. A pass means /llms.txt exists, is long enough, and is not disallowed. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_llms_txt_links
ChatGPTFetch the public website for one domain and run only this check: links in llms.txt on the same site respond. A pass means links in llms.txt on the same site respond. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_open_graph_tags
ChatGPTFetch the public website for one domain and run only this check: Open Graph title, description, and image are present. A pass means Open Graph title, description, and image are present. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_question_headings
ChatGPTFetch the public website for one domain and run only this check: question headings and FAQPage are present. A pass means question headings and FAQPage are present. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_robots_tag_agreement
ChatGPTFetch the public website for one domain and run only this check: meta robots and X-Robots-Tag agree on noindex. A pass means meta robots and X-Robots-Tag agree on noindex. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_single_h1
ChatGPTFetch the public website for one domain and run only this check: the page has exactly one H1. A pass means the page has exactly one H1. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_sitemap_freshness
ChatGPTFetch the public website for one domain and run only this check: the newest sitemap lastmod is within 90 days. A pass means the newest sitemap lastmod is within 90 days. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_sitemap_urls
ChatGPTFetch the public website for one domain and run only this check: a sitemap exists and its URLs respond. A pass means a sitemap exists and its URLs respond. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
site_check_title_and_description
ChatGPTFetch the public website for one domain and run only this check: title and description sit in the target length bands. A pass means title and description sit in the target length bands. No workspace and no other check. For the signed-in workspace audit, use analyze_site.
start_content_job
ChatGPTStart one draft job for a content Plan of Action step and a calendar day. The topic and language come from that step and the workspace brand language. Returns the job id and state. A repeated idempotency key returns the same id. This does not publish and does not return the article.
start_onboarding_v2_scan
ChatGPTStart an AI search scan for a workspace that is still in onboarding. It sends that workspace’s buyer prompts to the tracked AI engines and stores the answers. The default mode, warm, starts the scan. The generate mode closes the scan and starts the report timer, which finishes within 90 seconds. Requires the workspace id, the onboarding prompt batch id, and the workspaces:write permission.
submit_feedback
ChatGPTSend a note about MencionAI to the product team. Use when the user asks to give feedback or you see a concrete product improvement. Set anonymous true to hide the person. Does not list prior notes.
upload_blog_font
ChatGPTUpload a woff2, woff, ttf, or otf file (max 2MB) and use it on the customer blog. The file wins over a listed font. Send raw base64 or a data URL.
upload_blog_logo
ChatGPTUpload a png, jpeg, or webp logo (max 5MB) for the customer blog. Send raw base64 or a data URL.
validate_geo_content
ChatGPTCheck a customer blog draft against the same AEO/SEO/GEO/AIO citation contract used to publish first-party MencionAI pages. Returns missing_obligations. Call after prepare_geo_content_brief and writing. Related links must be locale-prefixed on the workspace domain.
visibility_citations
ChatGPTPower-user fallback — prefer get_visibility_overview for a short citation sample. Domains and URLs from the current Project visibility snapshot. Omit limit to return every source.
visibility_competitors
ChatGPTPower-user fallback — prefer compare_competitors for ranked brand vs rivals. Competitor hosts and mention/citation counts.
visibility_coverage
ChatGPTPower-user fallback — prefer get_visibility_overview or explain_visibility_gaps. Per-prompt mention coverage (mentioned/cited flags and counts).
visibility_mentions
ChatGPTPower-user fallback — prefer get_visibility_overview. Brand mentions from the current visibility snapshot, grouped by tracked prompt text.
visibility_responses
ChatGPTAI answer excerpts from the visibility snapshot. Requires API key scope visibility:answers:read. Fails closed if the scope is missing.
visibility_series
ChatGPTDaily AI/SEO/visibility score series for a workspace (range 7d, 30d, or 90d; default 30d).
visibility_summary
ChatGPTPower-user fallback — prefer get_visibility_overview. Current Project visibility snapshot summary (prompts, mentions, citations).
wait_for_ai_search_scan
ChatGPTStart an AI search mention scan for a workspace and wait until the job finishes or the timeout (default 60 seconds, maximum 120). Uses plan quota. Can refresh the visibility overview when the scan completes.
wait_for_onboarding_v2_report
ChatGPTWait until the onboarding AI search report for one prompt batch is ready or partial, or until the timeout. Default wait is 90 seconds and the maximum is 180 seconds.