85/100
Grade B · Ready for agents
Ahead of 51% of 75 sites scanned
Top fixes — ranked by impact
- 01Publish a sitemap.xml (or sitemap_index.xml) and reference it in robots.txt.
- 02Speed up server response (caching, CDN, server tuning) to get TTFB under ~0.8s.
- 03To raise content value: add data tables, datasets, or structured records — the most citable content for AI.
- Sitemaps No sitemap.xml found. A sitemap gives crawlers and agents a complete, machine-readable map of your pages so none are missed. fail +6 pts
-
Loading speed
Real-user server response (TTFB) is 876 ms at the 75th percentile — a little slow for crawlers.
Crawlers apply tight timeouts and crawl budgets — a slow server response gets fetched less often and sometimes abandoned mid-load. Real-user field data (Chrome UX Report) reflects actual server speed better than a single synthetic ping.
warn +3 pts
Details
- TTFB 876 ms (real-user)
- LCP 1.6 s
- INP 98 ms
- CLS 0.01
-
Content quality
Title, description, and image alt text: 67% ideal.
Concise titles, meta descriptions of the right length, and alt text on images give agents accurate, quotable summaries of each page and its media.
warn +2 pts
Details
- Title 54 chars
- Description 149 chars
- Image alt text: 4/35 descriptive · 31 empty alt=""
-
Content value
Content value: 77% — richest page: /api.
Whether an AI actually uses and cites a page depends on its content: structured, tabular, and data-rich content is the most citable, followed by fact-rich, answer-shaped, cleanly chunked text — plain prose is the least valuable.
warn +2 pts
Details
- Data richness: 2 data table(s), 108 data link(s)
- Substance: 996 words of main content
- Fact density (numbers, dates, entities)
- Answer-shaped structure (Q-headings, lists, definitions)
- Chunkability (heading hierarchy)
- Self-containedness (low link-density, no vague CTAs)
- Machine-readable data No obvious feeds or public APIs detected. Feeds and public APIs let agents consume your data structurally instead of scraping HTML. warn +1 pts
-
Content negotiation
No agent-friendly Markdown representation offered via content negotiation.
Serving a Markdown representation via Accept negotiation gives AI agents clean content without parsing HTML — an emerging, forward-looking signal.
warn
Details
- Serves Markdown for Accept: text/markdown
- Sets Vary: Accept
- Returns 406 for unsupported types (optional per RFC 9110)
- Honors q-values
-
Meta & semantics
5/6 meta/semantic signals present.
Titles, descriptions, canonical URLs, and a clean heading structure help agents identify, summarize, and attribute your pages correctly.
pass +1 pts
Details
- <title>
- meta description
- canonical link
- Open Graph tags
- html lang attribute
- exactly one <h1>
-
AI bot access
All AI crawlers are allowed.
AI assistants can only read, cite, or recommend your site if their crawlers aren't blocked in robots.txt. Search & answer crawlers are what get you cited; training crawlers only affect model corpora.
pass
Details
- OAI-SearchBot (AI search) allowed
- PerplexityBot (AI search) allowed
- Claude-SearchBot (AI search) allowed
- DuckAssistBot (AI search) allowed
- YouBot (AI search) allowed
- Amazonbot (AI search) allowed
- Applebot (AI search) allowed
- ChatGPT-User (user-triggered) allowed
- Perplexity-User (user-triggered) allowed
- Claude-User (user-triggered) allowed
- meta-externalfetcher (user-triggered) allowed
- GPTBot (model training) allowed
- ClaudeBot (model training) allowed
- anthropic-ai (model training) allowed
- Google-Extended (model training) allowed
- Applebot-Extended (model training) allowed
- CCBot (model training) allowed
- Bytespider (model training) allowed
- meta-externalagent (model training) allowed
- cohere-ai (model training) allowed
- PetalBot (model training) allowed
- Content in HTML 1618 words of server-rendered text in semantic <main>/<article>. Most AI crawlers don't execute JavaScript — content must be in the initial HTML or agents simply won't see it. pass
-
Structured data (JSON-LD)
Valid structured data: WebSite, Organization, Person, ImageObject, ContactPoint, Place, PostalAddress.
Schema.org JSON-LD tells agents exactly what your entities are (products, articles, your organization) so they can extract facts reliably instead of guessing. Incomplete or invalid markup — missing @context, or a Product with no price — is skipped just like missing markup.
pass
Details
- WebSite — add recommended: potentialAction
- Person — add recommended: url, jobTitle, sameAs, image
- Person — add recommended: url, jobTitle, sameAs, image
- Organization ✓ complete
- Bot reachability Site responds 200 to a bot user-agent. If your WAF or bot filter blocks crawler user-agents, agents can't fetch the page at all — regardless of how well it's built. pass
-
HTTPS & redirects
Served over HTTPS.
HTTPS is a baseline trust and ranking signal; if the insecure http:// version also serves content without redirecting, agents can index a duplicate, non-canonical copy.
pass
Details
- HTTPS served
- Indexing directives No indexing or AI opt-out directives block crawlers. Meta robots tags and X-Robots-Tag headers can quietly tell search and AI crawlers to skip your page — often left on by accident from a staging setup. pass
- llms.txt llms.txt is present. An llms.txt file gives AI agents a curated summary and key links, guiding them to your most important content. pass
Embed your badge
