Public AI readability report

crawlable.synthworx.com

Scanned Fri, 31 Jul 2026 16:41:41 GMT · https://crawlable.synthworx.com/repros/2026-07-31-wasm-table/index.html

70
out of 100

AI assistants cannot properly read this site

95% of your page content only exists after JavaScript runs

Access100/100
Can crawlers fetch it
Readability33/100
Is content in the HTML
Structure68/100
Machine-readable meaning
Answerability60/100
Shaped to be quoted

Send this to whoever can deploy the fix

We’ll email the full findings plus a prioritised fix checklist for crawlable.synthworx.com. Access failures come first, followed by response-content and structure fixes. No account needed.

Initial response and browser render

Word counts of readable text before and after browser execution. The gap is a rendering dependency, not a claim about every vendor’s crawler behavior.

Initial HTML response
16
words of readable text
Browser render after JavaScript
341
words of readable text

Crawler access

Live requests were sent with each vendor’s published user-agent string. This tests how the site handles the claimed identity; a user-agent alone does not authenticate the vendor. A robots.txt pass with a failed request points to another access control.

CrawlerVendorRolerobots.txtLive request
GPTBotOpenAITrainingAllowedHTTP 200
OAI-SearchBotOpenAISearch indexretrievalAllowedHTTP 200
ChatGPT-UserOpenAILive fetchretrievalAllowednot probed
ClaudeBotAnthropicTrainingAllowedHTTP 200
Claude-SearchBotAnthropicSearch indexretrievalAllowednot probed
Claude-UserAnthropicLive fetchretrievalAllowednot probed
PerplexityBotPerplexitySearch indexretrievalAllowedHTTP 200
Perplexity-UserPerplexityLive fetchretrievalAllowednot probed
Google-ExtendedGoogleTrainingAllowednot probed
Applebot-ExtendedAppleTrainingAllowednot probed
meta-externalagentMetaTrainingAllowednot probed
AmazonbotAmazonSearch indexAllowednot probed
BytespiderByteDanceTrainingAllowednot probed
CCBotCommon CrawlTrainingAllowednot probed

Findings (4)

1 critical access or response issue found in this scan.

critical

95% of your page content only exists after JavaScript runs

Evidence
Raw HTML contains 16 words. After rendering, the page contains 341 words.
Why it matters
No major AI crawler executes JavaScript — GPTBot, OAI-SearchBot, ClaudeBot, Claude-SearchBot and PerplexityBot all read raw HTML only. Google renders JS, so this site can rank normally in Google search while being close to blank in ChatGPT, Claude and Perplexity. That gap is invisible in Google Search Console.
Fix
Serve the primary content in the initial HTML response via server-side rendering, static generation, or prerendering for crawler user-agents. Verify with `curl` — if the content is not in the raw response, AI crawlers do not have it.
high

No structured data on the homepage

Evidence
No JSON-LD blocks and no microdata found.
Why it matters
Structured data is how a machine learns your entity name, what you sell, where you operate and how to contact you, without inferring it from prose. Without it, assistants are more likely to describe you vaguely or confuse you with a similarly named company.
Fix
Add JSON-LD for Organization (name, url, logo, sameAs, description) plus the type matching the page — Product, Service, LocalBusiness or Article.
medium

No question-shaped content

Evidence
No question headings and no FAQPage schema found.
Why it matters
Assistants answer questions. Pages organised as explicit question/answer pairs are markedly easier to retrieve and quote than undifferentiated marketing prose.
Fix
Add a server-rendered FAQ section using the real questions prospects ask, marked up with FAQPage schema.
low

Missing meta description

Evidence
No meta[name="description"] in the raw HTML.
Why it matters
Removes a concise, author-controlled summary that assistants frequently reuse verbatim.
Fix
Add a 140–160 character description stating what the business does.

Get these fixed and keep them fixed

The diagnosis stays free. Paid plans generate reviewable robots.txt, JSON-LD and llms.txt artifacts, then re-scan on schedule and email you when a check regresses.

See plansScan another site