Components
Total: 100 points. Components are weighted by their impact on AI crawler access. Some have circuit breakers — critical failures cap the maximum score regardless of other factors.
Crawler Access & Discovery covers more ground than it used to, but its share of the 100 points is unchanged: the newer site-wide checks were carved out of the existing budget rather than added on top, so scores stay comparable across the change.
Site-wide checks that have not run yet — on a brand new project, or a domain that could not be reached — are left out of the calculation entirely rather than counted as a pass or a failure. A score is only ever built from what was actually measured.
Score levels
Crawler access & discovery
zued checks access individually for each AI crawler: GPTBot, ClaudeBot, PerplexityBot, Bingbot, Google-Extended, and others. A singlerobots.txt rule or meta robots tag can block one engine while leaving others unaffected.
robots.txt is what your site declares. What your server and its bot protection actually do when an AI crawler asks for a page is a separate question, and so is whether your site has declared that its content should not be used inside AI answers. Both are checked site-wide and both feed this component. See AI crawler access for what each check means and why one of them deliberately costs you nothing.
Your XML sitemap is checked too. It is how engines discover pages you never linked prominently, so a missing or empty sitemap limits how much of your site ever enters the candidate pool.
JS rendering
If your content requires JavaScript to appear, some AI crawlers may never see it. zued compares your page before and after JavaScript execution and flags content that only appears post-render as a crawlability risk.Server response time (TTFB)
Time to First Byte measures how quickly your server begins responding to a request. For AI visibility, this matters at the retrieval stage: when an AI engine’s crawler requests your page, a slow server response reduces the number of pages it can fetch within its crawl budget and may cause it to deprioritize your domain entirely. Pages with consistently slow server response times are less likely to enter the candidate pool that AI engines draw from when generating answers. zued measures TTFB and applies score caps when response times are high enough to impact crawl efficiency. Target response times:
Google recommends a server response time under 200ms. For AI crawlers, this target is even more important: LLM retrieval operates under strict latency budgets during real-time search, and sites with fast response times receive significantly more crawler requests than slower sites.
Structured data
Structured data (Schema.org) is confirmed to assist Copilot’s understanding. Its direct benefit for ChatGPT, Perplexity, and Gemini citation behavior is not established. zued scores its presence accurately for what it affects.