live_checks_12082026_0223.md
/app/data/llm/analysis/bupa/live_checks_12082026_0223.md
All 21 probes returned 200, with zero blocks, zero errors, and no differential between browser and bot responses. Every AI crawler tested — Bytespider, ClaudeBot, GPTBot, Google-Extended, OAI-SearchBot, PerplexityBot — accessed all three paths without any sign of blocking or rate-limiting. The site is technically open to AI crawlers, at least from this probe source.
What stands out
- No blocking at all. Across 6 AI-specific user agents and 3 representative paths, not a single 403, 429, or captcha. This is a clean sheet. If the site uses Cloudflare, the lack of blocking from a datacenter IP is notable — it means the WAF is not aggressing on these bot signatures, or that the probe IP is not triggering any bot mitigation rules. Either way, the path is clear.
- Response times are healthy. AI bots saw 42–84 ms average latency, with p95s under 30 ms for most. The browser agent hit 174 ms average, driven entirely by a 477 ms outlier on robots.txt. That outlier is almost certainly a one-off network delay or a larger robots.txt file, not a systemic issue. For crawlers, the experience is fast and consistent.
- robots.txt is accessible but its directives are unknown. The probe returns 200, confirming the file exists and is fetchable. However, the content is not included in this dataset. If the file contains
Disallowrules for any of these bots, the 200s on the pages would still be returned — only compliant crawlers would obey. So the HTTP-level accessibility is unconditional; crawlability depends on robots.txt content, which must be checked separately.
What's positive
- The site presents zero technical resistance to AI crawlers. This is a prerequisite for appearing in AI-generated search results. If the site is not being cited or surfaced in AI overviews, the cause is not here — no blocking, no cloaking, no errors.
- The uniformity of responses across all agents (identical status, similar timing, no differential content serving) suggests a single, stable serving stack. No crawler-specific treatment is evident, which reduces the risk of accidental exclusion.
What to verify (hypotheses and limitations)
- robots.txt content. Fetch the file and check for
Disallowdirectives targeting any of these user agents. The 200 status alone is not enough to confirm crawlability. - Single IP source. The probes originate from one datacenter IP. Cloudflare's bot management can behave differently for residential vs. datacenter traffic, and for different IP reputation scores. A 0% block rate from this IP does not guarantee that all production crawlers are unblocked everywhere, but it strongly suggests no blanket blocking of these user agents.
- Crawlability vs. indexability. No blocking is necessary but not sufficient for AI visibility. The actual status of AI citations (e.g., in ChatGPT, Perplexity) will depend on whether these bots are actively crawling, how the content is parsed, and how the site's authority is evaluated. This data only says the door is open.
Potential featured finding
The fully open door for AI crawlers is the headline. If combined with other signals — e.g., low AI citation counts despite strong search rankings — this finding would point the investigation elsewhere: content quality, structured data, brand authority in AI training data, or the site's crawl frequency. Conversely, if the site is performing well in AI search, this data confirms that no technical blocker is holding it back.
Key signals
- 0% block rate across all AI bots and paths — no Cloudflare-imposed restrictions, captchas, or access controls from this probe source.
- All 21 probes returned 200 with no errors and no differential between browser and bot responses, ruling out cloaking or access gating.
- AI bot response times are consistently fast (42–84 ms avg), ensuring crawl efficiency is not a bottleneck.
- robots.txt accessibility confirmed (200), but its directives are unverified — a critical gap in determining actual crawlability for compliant bots.
- The single-source IP limitation means production blocking cannot be ruled out entirely, but the uniform openness is a strong positive indicator.