Does your page have all four? SEO, GEO, AEO, and AIO — 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, complete share metadata, alt attributes on images, and useful internal links. 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 simply rank pages — they synthesize answers from sources they judge useful and credible. So the job shifts from “rank on this SERP” to “be one of the sources worth citing when a real buyer asks a real question.” This pillar scores substantive depth, identifiable author evidence, and outbound references to support the claims being made. It is a coarse indicator based on three independently scored signals, so it moves in larger steps than the other pillars. Freshness is shown as context but is not scored, because many strong pages are legitimately evergreen.

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.

AIO asks whether AI systems can reach, discover, and parse the page at all. It is the technical access layer beneath GEO and AEO: AI crawler rules in robots.txt, machine-readable entity schema, a valid llms.txt, sitemap discovery, and a small presence check for the experimental Open Knowledge Format. AIO is separate from GEO because access is not quality. A page can be perfectly crawlable and still offer nothing worth citing; it can also be excellent source material that a blocked crawler cannot read. When an external AIO check fails, the checker labels it unknown and removes those points from the denominator rather than treating its own fetch failure as the site’s failure.

The tool fetches your page server-side and evaluates distinct signals across four independent pillars. Here is the breakdown:

Each scored signal belongs to one pillar only. llms.txt, AI crawler access, sitemap discovery, and machine-readable entity schema now count under AIO. OKF presence is binary and contributes at most 2 of AIO’s 47 raw points. Content freshness remains informational under GEO. If an external AIO fetch fails, that signal is labeled unknown and excluded from the denominator. This scoring model changed when AIO became a fourth pillar, so pages checked under the earlier three-pillar model will score differently now.

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

• 90–100 (grade A). Elite. This page is doing the job across all four pillars.
• 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 rankings, citations, answer placements, or AI access 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.
• External AIO checks can be unknown. A timeout, server error, authorization response, or failed request is excluded from the AIO denominator and labeled unknown. A confirmed missing file still scores as absent.
• OKF is a presence check only. The checker does not validate the bundle’s quality or usefulness, so OKF is capped at 2 raw AIO points.
• 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.

AIO combines established access controls with newer discovery formats. The checker reads robots.txt for GPTBot, ClaudeBot, PerplexityBot, and Google-Extended; checks sitemap discovery; and measures machine-readable entity schema in the page. It also validates llms.txt, a Markdown map at /llms.txt that gives AI systems a curated route to important pages. Open Knowledge Format (OKF) is a Google Cloud v0.1 format for agent-readable knowledge bundles at /okf/. Because this checker detects only whether an OKF path exists, not whether the bundle is valid or useful, it contributes no more than 2 raw AIO points. Read the full llms.txt guide →