Does your page have all three? SEO, GEO, and AEO — scored in seconds.

SEO is the discipline of getting a page to rank on a traditional search results page, principally Google. The classic on-page signals still matter: title tag inside the 25–60 character range, meta description 70–160 characters, exactly one clean H1, canonical URL, HTTPS, a mobile viewport meta tag, Open Graph metadata, alt text on every image, and enough content depth to compete for the query. Off-page, backlinks and brand authority still tilt the ranking, but the ten blue links are no longer the whole search page — AI Overviews, People Also Ask, and featured snippets sit above them. Google’s own SEO Starter Guide is still the canonical reference for the fundamentals.

GEO is the craft of getting your content cited when large language models like ChatGPT, Claude, Perplexity, and Google Gemini synthesize an answer. LLMs do not rank pages — they synthesize answers from trusted sources. So the job shifts from “rank on this SERP” to “be one of the sources the LLM trusts enough to quote when a real buyer asks a real question.” What LLMs reward: content depth (1,200+ words for competitive queries), question-format headings they can chunk cleanly into Q&A pairs, JSON-LD structured data declaring your entities per the Schema.org vocabulary, author byline plus linked Person schema, and outbound citations to authoritative sources (.gov, .edu, .org, primary research).

AEO is the craft of winning the boxed answer at the top of a search result, or the equivalent inside an AI response. That includes Google’s featured snippet, AI Overview citations, People Also Ask expansions, and the direct answer inside a ChatGPT, Claude, or Perplexity response. All of these reward the same page structure: a question in the heading, a direct 40–60 word answer in the first sentence of the paragraph immediately underneath, then supporting detail. The tactical wins are FAQPage schema, HowTo schema for step-based content, question-driven H2s, answer-first paragraphs, and tables for comparisons.

The tool fetches your page server-side and evaluates dozens of signals across three independent pillars, plus emerging readiness checks. Here is the breakdown:

Plus emerging signals: llms.txt validation (Anthropic-proposed Markdown map for LLMs — see the llmstxt.org spec), Open Knowledge Format bundle detection, AI crawler access in robots.txt (GPTBot, ClaudeBot, PerplexityBot, Google-Extended), sitemap presence, and content freshness. Emerging signals are informational — they do not count against the trifecta score.

Each pillar produces a 0–100 score and a letter grade. The overall score is the average of the three pillars. Here is how to interpret what you get back:

• 90–100 (grade A). Elite. This page is doing the job across all three surfaces.
• 80–89 (grade B). Solid. There are a few specific fixes but no fundamentals are broken.
• 70–79 (grade C). Real headroom. Usually one or two structural fixes will move it 10+ points.
• 60–69 (grade D). Weak. Multiple failures compounding. Prioritize schema and question-format headings.
• Under 60 (grade F). Broken. The page is losing citations, rankings, or snippets that it should be winning.

The tool tells you exactly which items passed, which failed, and which are informational notes. Fix the failures top-down — the list is ordered by impact.

• Competitiveness is not authority. The checker scores how an answer page is constructed, not how trusted its source already is. Wikipedia, .gov, .edu, and major news sites can score low because they do not need FAQPage schema to earn citations. If your site is not already one of those authorities, these are still useful signals to improve.
• It is a static-HTML scorer. It fetches server-rendered HTML. Fully client-rendered React or SPA pages can look nearly empty and score badly even when they work for users. Server-rendered and statically generated pages are represented correctly.
• Word-count and citation thresholds are heuristics. A short, excellent page can beat a long one. “Add more words” is a signal, not a mandate.
• It does not test rendering, Core Web Vitals, backlinks, or whether the content is actually good. Those require browser, authority, and editorial analysis beyond a static HTML check.
• It reports what it can see in the HTML. It cannot tell you whether a page deserves to rank.

Two new standards are worth being early on. llms.txt is a Markdown map served at the root of your site (/llms.txt) that gives large language models a curated list of your most important pages. The proposal came from Anthropic, and Anthropic itself, Vercel, and Cloudflare have already shipped one. The checker validates yours against the format spec. Open Knowledge Format (OKF) is a Google Cloud v0.1 spec for agent-readable knowledge bundles at /okf/ — informational only in the checker, because it is not a ranking signal, but forward-looking readiness worth knowing about. Neither is table stakes yet, but early adoption has real upside if either becomes a weighted signal in 2027. Read the full llms.txt guide →