Two Webs

Publishers / si.com

90

Mostly one web

11 of 14 AI agents get the full page at https://si.com/. Checked 2026-09-25 09:33 UTC.

si.com scores 90 out of 100 for machine visibility, number 7 of the 100 content publishers checked on 2026-09-25. 11 of 14 AI agents receive the full page: 5 of 8 training crawlers, 3 of 3 AI search indexes and 3 of 3 assistant fetchers. All 3 refusals happen at the HTTP level with no matching robots.txt rule, so a crawler cannot tell the block is deliberate. The pattern is "cite me, don't train on me": 3 of 8 training crawlers are refused while 6 of 6 search and assistant fetchers get through. The site publishes an llms.txt.

youfull pagedegradedblockedunknown

Human

Status
200 → https://www.si.com/
Title
Sports Illustrated
H1
none
Words
30,790
Scripts
120 tags · 87 KB
Meta robots
max-image-preview:large
Canonical
https://www.si.com/
Server
CloudFront

Machine

robots.txt
20 groups · 3 sitemaps
llms.txt
present
llms-full.txt
missing
X-Robots-Tag
none
Blocked
Meta-ExternalAgent, Amazonbot, Bytespider
Degraded
none
Challenged
none

Gaps

  • medium
    3 of 14 AI agents are blockedMeta-ExternalAgent: HTTP 403 for this user agent · Amazonbot: HTTP 403 for this user agent · Bytespider: HTTP 403 for this user agent
  • info
    Training blocked, citation allowedTraining crawlers are refused while search and assistant fetches get through. This is the common "cite me, don't train on me" posture.
  • info
    Google-Extended not listedYou block other training crawlers but Gemini training via Google-Extended is still allowed. Add it to the same group if the policy is meant to be even-handed.
  • info
    Canonical points elsewhereCanonical is https://www.si.com/. Indexes will credit that URL, not this one.

Every agent

AgentTyperobots.txtFetchWordsResult
Human (Chrome)YouHumanallowedAllow: /200870 ms30,790visible
HTTP 200, 30790 words
GooglebotGoogle · Classic search index. Also feeds AI Overviews.Search engineallowedAllow: /2001022 ms30,790visible
HTTP 200, 30790 words
Google-ExtendedGoogle · Robots-only token controlling Gemini training use. Never fetches on its own.AI trainingallowedAllow: /n/avisible
Robots-only token; allowed by robots.txt
GPTBotOpenAI · Training crawler for OpenAI models.AI trainingallowedAllow: /2001044 ms30,790visible
HTTP 200, 30790 words
OAI-SearchBotOpenAI · Indexes for ChatGPT search results and citations.AI search indexallowedAllow: /2001602 ms30,790visible
HTTP 200, 30790 words
ChatGPT-UserOpenAI · Live fetch when a user asks ChatGPT about a page.AI assistant (live fetch)allowedAllow: /2001020 ms30,790visible
HTTP 200, 30790 words
ClaudeBotAnthropic · Training crawler for Claude models.AI trainingallowedAllow: /200895 ms30,790visible
HTTP 200, 30790 words
Claude-SearchBotAnthropic · Indexes for Claude web search.AI search indexallowedAllow: /200922 ms30,790visible
HTTP 200, 30790 words
Claude-UserAnthropic · Live fetch when a user asks Claude about a page.AI assistant (live fetch)allowedAllow: /200958 ms30,790visible
HTTP 200, 30790 words
PerplexityBotPerplexityAI search indexallowedAllow: /2003016 ms30,790visible
HTTP 200, 30790 words
Perplexity-UserPerplexityAI assistant (live fetch)allowedAllow: /2001727 ms30,790visible
HTTP 200, 30790 words
BingbotMicrosoft · Bing index. Also feeds Copilot and ChatGPT search.Search engineallowedAllow: /2001116 ms30,790visible
HTTP 200, 30790 words
Applebot-ExtendedApple · Robots-only token for Apple Intelligence training.AI trainingallowedAllow: /n/avisible
Robots-only token; allowed by robots.txt
Meta-ExternalAgentMetaAI trainingallowedAllow: /403758 ms0blocked
HTTP 403 for this user agent
AmazonbotAmazon · Alexa and Amazon AI.AI trainingallowedAllow: /403819 ms0blocked
HTTP 403 for this user agent
BytespiderByteDance · Known for ignoring robots.txt.AI trainingallowedAllow: /403818 ms0blocked
HTTP 403 for this user agent
CCBotCommon Crawl · Open dataset used to train most LLMs.AI trainingallowedAllow: /2001064 ms30,790visible
HTTP 200, 30790 words

Snapshot

Measured 2026-09-25 09:33 UTC from a residential connection. Sites change their rules often, and sites that verify crawler IPs answer differently to different networks. Run a live comparison to see what si.com does right now, from Cloudflare's network.

← substack.com (92)All 100 publisherscbsnews.com (89) →