audit_page
ClaudeFetches one public web page and runs basic on-page SEO and GEO (generative engine optimization) checks on its HTML: status and redirects, title, meta description, headings, canonical, indexability, language, mobile viewport, images without alt text, word count, Open Graph, structured data, answer-ready content signals, and whether major AI crawlers may access the page according to robots.txt. Returns a 0-100 score, the extracted page data and a prioritized list of findings with suggested fixes. Does not run JavaScript. Checks follow Google Search Central documentation (https://developers.google.com/search/docs) and the Robots Exclusion Protocol (RFC 9309, https://www.rfc-editor.org/rfc/rfc9309).
check_ai_crawler_access
ClaudeReads a website's robots.txt and reports, for 16 search and AI crawlers (GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, PerplexityBot, Googlebot, Google-Extended, Bingbot, Applebot and others), whether each may fetch a given path and which rule decides it. Also checks that the declared XML sitemap is reachable and whether an llms.txt file exists and follows the expected format. Rules are evaluated per the Robots Exclusion Protocol (RFC 9309, https://www.rfc-editor.org/rfc/rfc9309); sitemaps per https://www.sitemaps.org/protocol.html; llms.txt per https://llmstxt.org.
check_local_seo
ClaudeChecks the local SEO signals of a public page for a business with a physical location: name, address and phone (NAP) visible on the page and declared in LocalBusiness or Organization structured data, whether the two match, click-to-call phone links, map embed or link, visible opening hours, links to the Google Business Profile, and the city in title and H1. If the page has no business markup, the home page markup is also read. Optionally compares everything with the name, phone and address the user expects. Structured data per https://schema.org/LocalBusiness and Google local business guidelines (https://developers.google.com/search/docs/appearance/structured-data/local-business).
check_page_weight
ClaudeInspects the HTML of one public page for common loading-performance problems without rendering it: HTML size and compression, caching header, render-blocking scripts and stylesheets in the head, images without width and height, lazy loading (including a lazy-loaded first image), legacy image formats versus WebP/AVIF, responsive images, third-party script hosts, embedded iframes, and the file size of up to 8 images. It is not a substitute for Core Web Vitals field data. Recommendations follow web.dev performance guidance (https://web.dev/learn/performance) and Google Search Central (https://developers.google.com/search/docs/appearance/core-web-vitals).
compare_pages
ClaudeRuns the same on-page SEO and GEO checks on two public pages, typically the user's page and a competitor's page for the same search query, and returns them side by side: score, response time, title and description length, headings, amount of text, structured data types, FAQ markup, Open Graph, internal links, images without alt text and AI crawler access. Lists where the first page is behind and where it is ahead. Does not run JavaScript. Checks follow Google Search Central documentation (https://developers.google.com/search/docs).
crawl_site
ClaudeCrawls a public website from a start URL by following internal links (up to max_pages HTML pages, same host, obeying robots.txt) and reports site-wide issues: broken internal links with the pages that link to them, internal links that redirect, duplicate or missing titles and meta descriptions, missing or multiple H1, noindex pages, canonicals pointing elsewhere, thin and slow pages, link depth, and differences between the XML sitemap and the pages found. Returns a compact row per page and grouped findings. Does not run JavaScript, skips URLs with query strings and does not check external links. Crawling follows the Robots Exclusion Protocol (RFC 9309, https://www.rfc-editor.org/rfc/rfc9309) and Google Search Central guidance (https://developers.google.com/search/docs).
draft_llms_txt
ClaudeBuilds a draft llms.txt file (the llmstxt.org format: site name, short summary, sections of annotated links) for a public website. Reads the XML sitemap, or the home page links when there is no sitemap, fetches up to max_pages pages to take their titles and descriptions, and groups them by URL section. Returns the file text for the user to review; nothing is published. Format specification: https://llmstxt.org.
generate_schema_markup
ClaudeGenerates ready-to-paste JSON-LD (schema.org) markup from details the user provides: Organization, LocalBusiness (or a subtype such as Dentist or Restaurant), Person, WebSite or FAQPage. Returns the JSON-LD object, the HTML script tag and the recommended properties still missing. Works offline: it does not fetch any page. Output follows the schema.org vocabulary (https://schema.org) in JSON-LD (https://json-ld.org).
inspect_structured_data
ClaudeFetches one public web page, extracts its JSON-LD structured data (schema.org), reports JSON syntax errors, lists the entity types found and, for common types (Organization, LocalBusiness and subtypes, WebSite, Article, Product, FAQPage, BreadcrumbList, Person, Service, Event, HowTo, Review), the recommended properties that are missing. Also counts microdata items. Types and properties follow the schema.org vocabulary (https://schema.org) and Google structured data guidelines (https://developers.google.com/search/docs/appearance/structured-data).
request_expert_review
ClaudeSends the user's request for a manual SEO and GEO review of their website to Orobici (seoperaiexpert.it), the SEO agency in Bergamo, Italy, that runs this connector. Orobici receives the website, the email address and the optional name, company, phone and message by email and replies to that address. Use only when the user has asked to be contacted by Orobici and has confirmed the details to share; set user_consent to true only in that case.