3 tools expose this connector's capabilities to your AI agent.
You must address: (1) SPAMBRAIN COMPLIANCE — no keyword stuffing (density <1-2%), natural editorially-earned links, no scaled content abuse (each page must serve unique user intent), no cloaking/hidden text. SpamBrain is ML-powered and detects patterns, not rules — you cannot game it. (2) E-E-A-T SIGNALS — DEMONSTRATE (not claim) Experience, Expertise, Authoritativeness, Trustworthiness through author credentials, original data, external citations, transparency. (3) TECHNICAL FOUNDATION — self-referencing canonical tags (absolute URLs), Schema/JSON-LD (@graph + @id entity linking), Core Web Vitals (LCP <2.5s, INP <200ms, CLS <0.1), XML sitemaps, HTTPS, mobile-first responsive design. (4) GEO OPTIMIZATION — structure for AI citation: statistics every 150-200 words, source citations, expert quotes, modular autonomous sections, question-based headers. Target Citation Share as primary metric (Princeton/Georgia Tech research: 30-40% citation boost). (5) AEO STRUCTURE — answer-first format (40-60 word atomic answers after H2), entity consistency, FAQPage schema, conversational tone, multimodal (alt text, transcripts), extractability. If the tool rejects, your strategy has a critical gap. Fix it before publishing.
Ultra-expert SEO validation tool for the post-AI search era (2026+). Validates content strategy against Google SpamBrain ML detection, E-E-A-T quality framework, technical SEO fundamentals (Core Web Vitals INP, Schema JSON-LD, canonicalization), Generative Engine Optimization (GEO) for AI citation, and Answer Engine Optimization (AEO) for AI-powered answer engines. Built from the Google May 2026 Core Update, March 2026 SpamBrain Update, Princeton/Georgia Tech KDD 2024 GEO research, and Google AI Overviews impact data. Catches SpamBrain Trigger (keyword density that trips ML detection — page targets "best CRM software." The phrase appears 23 times in 1,200 words = 4.2% density. SpamBrain March 2026 update processes this in < 20 hours. Result: page drops from position 8 to position 47 overnight. The writer thought "more mentions = more relevance." The opposite is true since 2019. Target: < 1-2% density. Use semantic variations, related entities, natural language. SpamBrain detects PATTERNS, not individual signals — keyword density is just one trigger), E-E-A-T Theater (claiming expertise without demonstrating it — "Written by an expert with 10 years of experience." No author bio page. No LinkedIn profile linked via sameAs. No external citations. No original data. No verifiable credentials. Google Quality Raters evaluate: "Could I trust medical advice from this author?" If the answer requires clicking 3 links to verify: E-E-A-T fails. Fix: author schema with sameAs (LinkedIn, Twitter), published credentials on /about, original research cited by external sites, transparent editorial policy. DEMONSTRATE expertise through content quality and verifiable identity — not claims), Technical Decay (canonical pointing to a redirect or 404 — <link rel="canonical" href="https://example.com/old-url" /> — but /old-url 301s to /new-url. Google follows the canonical to /old-url, gets redirected, sees /new-url, but the canonical says /old-url. Conflict. Google ignores the canonical entirely. Now Google decides which URL to index — and it may choose the wrong one. Fix: canonical must point to the FINAL, indexable URL. Self-referencing. Absolute URL. Consistent between canonical tag, internal links, and XML sitemap entry. One URL. Everywhere. No exceptions), GEO Invisibility (content that AI engines cannot extract or cite — a 2,500-word article with zero statistics, zero source citations, zero expert quotes. Written as narrative prose — no question-based headers, no modular sections. When Perplexity/ChatGPT/Google AI Overview needs an answer about this topic: they cite the competitor whose content has: "According to McKinsey (2024), 67% of..." with clear H2 questions, 40-60 word atomic answers, and structured data. Princeton/Georgia Tech KDD 2024: structured content gets 30-40% higher citation rate. Fix: statistics every 150-200 words, source citations, modular autonomous sections, question-based H2/H3 headers, RAG-extractable formatting), and AEO Mismatch (wrong content format for answer engine extraction — target keyword: "how to set up a CRM." Intent: step-by-step guide. Your content: a 3,000-word essay about CRM philosophy. No numbered steps. No HowTo schema. Google AI Overview shows: competitor's 8-step guide with HowTo structured data. Your comprehensive essay is invisible to answer engines because it is not structured for extraction. AEO requires: answer-first format (40-60 word direct answer after H2), FAQPage schema for Q&A content, HowTo schema for procedural content, entity consistency across all platforms). Call once per page/site strategy requiring search excellence
Address ALL: (1) TITLE: exact text, 50-60 chars, <580px, unique, front-loaded keywords, aligned with H1. (2) META DESCRIPTION: 140-155 chars, front-load first 120 chars, include CTA. (3) H1: single per page, thematically consistent with title (not identical). (4) HEADING HIERARCHY: H1→H2→H3→H4, never skip levels, describe content not style. (5) SEMANTIC: header, nav, main, article, section, aside, footer. No div soup. (6) DEPRECATED: zero usage of center, font, marquee, blink, frame, big, acronym, strike. (7) META TAGS: charset=UTF-8, viewport, lang, robots directives. (8) OPEN GRAPH: og:title, og:description, og:image (1200x630), og:url, og:type. (9) IMAGES: alt 80-125 chars, width/height mandatory, never lazy-load LCP, AVIF/WebP. (10) CANONICAL: self-referencing, absolute URLs, consistent with sitemaps/links.
Validates HTML for semantic perfection: title tag (50-60 chars/<580px), meta description (140-155 chars), single H1 aligned with title, heading hierarchy (H1→H2→H3, no skipping), HTML5 semantic elements, zero deprecated tags, Open Graph, image optimization (alt/dimensions/lazy-load strategy), and canonicalization. Built from current web standards and Google 2026 guidelines. Catches Title Truncation (title tag too long — gets cut in SERP — "Best Project Management Software for Small Teams - Comprehensive Guide 2024" = 72 chars. Google truncates at ~580px (~60 chars): "Best Project Management Software for Small..." The most important differentiator ("Comprehensive Guide 2024") is invisible in search results. Users see a generic, incomplete title and skip to the next result. CTR drops 23%. Fix: front-load keywords and value in first 50 chars. "Best PM Software for Small Teams (2024 Guide)" = 47 chars. Nothing truncated), Meta Description Duplication (same description across hundreds of pages — an e-commerce site uses the same template description across 500 product pages: "Find the best products at great prices. Shop now!" Google Search Console: "Duplicate meta descriptions: 487 pages." Google ignores all 487 descriptions and auto-generates snippets from page content instead — often pulling irrelevant text (navigation links, footer text, cookie notices). Fix: unique description per page. Each description: 140-155 chars, value front-loaded, CTA for transactional pages. Programmatic generation for product pages using product attributes), Multiple H1 Tags (confusing heading hierarchy — page has 3 H1 tags: the logo, the hero text, and the page title. Google's heading interpretation: "this page has 3 equally important topics." The actual page topic (page title) is diluted by the logo H1 and hero H1. Screen readers announce all 3 as top-level headings — confusing for accessibility. Fix: single H1 per page = the page title. Logo: link, not heading. Hero text: H2 or <p>), Deprecated Elements (legacy HTML tags triggering validator errors — <center>, <font color="red">, <marquee> — all deprecated in HTML5. Validators flag them. Screen readers may misinterpret them. Search engines may ignore styled content. Common in legacy codebases: <b> instead of <strong> (semantic difference — <strong> means "important," <b> means "bold" with no semantic meaning). Fix: replace with CSS classes. <center> → text-align:center. <font> → CSS color/font-family), and LCP Image Lazy-Load (hero image with loading="lazy" destroying Core Web Vitals — the hero image is the Largest Contentful Paint element. loading="lazy" tells the browser: "don't load this until it's near the viewport." But it IS in the viewport on page load. The browser waits. LCP: 4.2 seconds (fail: > 2.5s). Fix: hero/LCP image must have loading="eager" + fetchpriority="high". lazy-load ONLY for below-the-fold images (product gallery, blog images). Also: width/height attributes mandatory on ALL images — prevents CLS (layout shift)). Call once per page or template requiring HTML semantic validation
Address ALL: (1) SCHEMA TYPES: Organization, WebSite+SearchAction, BreadcrumbList, content type. (2) CONTENT ALIGNMENT: schema MUST match visible content — author, dates, prices, titles. Any mismatch triggers manual actions and AI distrust. (3) ENTITY LINKING: @graph to combine types, @id as unique identifiers, sameAs for profiles. (4) AI OPTIMIZATION: FAQPage (LLMs prefer Q&A), HowTo (step-by-step), speakable, SearchAction, dateModified (AI recency bias). PerplexityBot/OAI-SearchBot process JSON-LD. (5) VALIDATION: Google Rich Results Test + Schema Markup Validator. (6) RECENCY: dateModified current, review quarterly, stale = lower AI citation priority.
Validates JSON-LD structured data for AI engine optimization. Ensures schema types are present, accurately match visible page content, use @graph/@id entity linking, include AI-critical types (FAQPage, HowTo, speakable), and pass validation. Targets Google Gemini AI Overviews, Perplexity (PerplexityBot), ChatGPT (OAI-SearchBot), and Google (GoogleOther). Built from 2026 research on how AI crawlers process structured data to verify facts and build responses. Catches Schema Absence (zero JSON-LD on page — complete invisibility to structured search — the page has excellent content but no structured data. Google cannot verify: who wrote it (no Person schema), when it was updated (no dateModified), what organization publishes it (no Organization schema). Result: no rich results (no author photo, no star ratings, no FAQ accordion). No knowledge graph entry. AI engines cannot cross-reference author identity. Fix: minimum JSON-LD: Organization, WebSite with SearchAction, BreadcrumbList, content type), Content Mismatch (schema says one thing, page shows another — JSON-LD: "author": {"name": "Staff Writer"}. Visible byline on page: "Written by Jane Chen, CTO." Google Quality Raters find: schema author ≠ visible author. Result: manual action risk (hidden markup violation) + AI distrust (Perplexity cites "Staff Writer" instead of the authoritative "Jane Chen, CTO"). Every schema property MUST match visible content: author, dates, prices, ratings, titles), Entity Fragmentation (no @graph or @id linking — entities are isolated — the page has 4 separate <script type="application/ld+json"> blocks. Person, Organization, Article, BreadcrumbList — each standalone. Google cannot determine: does this Person AUTHOR this Article? Does this Organization PUBLISH this Article? Is this BreadcrumbList FOR this Article? Fix: single @graph array with @id references: "author": {"@id": "/#author"} links to Person with "@id": "/#author". This builds knowledge graph connections — isolated blocks build nothing), Stale dateModified (last updated 18 months ago — AI engines deprioritize — dateModified: "2024-06-15". Current date: 2026-01-20. Content is 18 months stale. AI engines have STRONG recency bias: Perplexity and Google AI Overviews prefer content updated within the last 6 months. Result: AI cites the competitor who updated their equivalent page last month. Fix: review content quarterly. Update dateModified only when content genuinely changes. Fake dateModified updates (changing the date without changing content) is a SpamBrain signal), and Validation Failure (malformed JSON-LD that Google cannot parse — trailing comma in JSON array: "sameAs": ["https://linkedin.com/in/author",]. Missing closing bracket. Wrong @type value ("Articles" instead of "Article"). Google Rich Results Test: "Error: Invalid JSON." No rich results. No structured data processing. Invisible to AI engines. Fix: validate with Google Rich Results Test AND Schema Markup Validator before publishing. Both tools. Every page. Zero errors). Call once per page requiring JSON-LD structured data validation