Weekly report · 2002-W21

The state of AI crawling

Measured on 30,030 reachable domains for the week ending 2002-05-26. Every figure is the result of an automated audit run the same way on every domain: read robots.txt, resolve the rules for 48 tracked agents, then send live requests as four of them.

← 2002-W20  ·  2002-W22 →

30,030
domains measured
8%
block an answer engine
19%
refuse AI agents at the edge
15%
publish an llms.txt

Block rate by crawler with change on the previous week

CrawlerOperatorContent used forBlock rateSitesWeek on week
GPTBot OpenAI training 8.2% 2,447 of 30,024 new
CCBot Common Crawl Foundation archive 7.9% 2,363 of 30,024 new
Bytespider ByteDance training 7.6% 2,291 of 30,024 new
ClaudeBot Anthropic training 6.9% 2,062 of 30,024 new
Meta-ExternalAgent Meta training 6.0% 1,795 of 30,024 new
Applebot-Extended Apple training 5.9% 1,766 of 30,024 new
Amazonbot Amazon training 5.9% 1,762 of 30,024 new
Google-Extended Google training 5.9% 1,761 of 30,024 new
anthropic-ai Anthropic training 5.6% 1,675 of 30,024 new
omgilibot Webz.io training 5.5% 1,656 of 30,024 new
Diffbot Diffbot training 5.4% 1,610 of 30,024 new
cohere-ai Cohere live retrieval 5.3% 1,593 of 30,024 new
omgili Webz.io training 5.1% 1,542 of 30,024 new
SemrushBot-OCOB Semrush training 3.5% 1,054 of 30,024 new
PerplexityBot Perplexity answer index 3.4% 1,017 of 30,024 new
ChatGPT-User OpenAI live retrieval 3.3% 997 of 30,024 new
FacebookBot Meta training 3.1% 938 of 30,024 new
YouBot You.com answer index 3.0% 893 of 30,024 new
Timpibot Timpi training 2.7% 825 of 30,024 new
Google-CloudVertexBot Google live retrieval 2.7% 824 of 30,024 new
ImagesiftBot ImageSift (Hive) training 2.7% 821 of 30,024 new
GoogleOther Google training 2.6% 785 of 30,024 new
Scrapy Zyte (open-source framework) training 2.4% 725 of 30,024 new
AI2Bot Allen Institute for AI training 2.3% 693 of 30,024 new
OAI-SearchBot OpenAI answer index 2.0% 615 of 30,024 new
cohere-training-data-crawler Cohere training 2.0% 609 of 30,024 new
MistralAI-Training Mistral AI training 1.9% 578 of 30,024 new
Meta-ExternalFetcher Meta live retrieval 1.8% 535 of 30,024 new
PanguBot Huawei training 1.7% 513 of 30,024 new
DuckAssistBot DuckDuckGo live retrieval 1.7% 510 of 30,024 new
Claude-User Anthropic live retrieval 1.6% 490 of 30,024 new
Perplexity-User Perplexity live retrieval 1.6% 469 of 30,024 new
Claude-SearchBot Anthropic answer index 1.5% 459 of 30,024 new
MistralAI-User Mistral AI live retrieval 1.5% 453 of 30,024 new
Kangaroo Bot Kangaroo LLM training 1.5% 447 of 30,024 new
Meta-WebIndexer Meta answer index 1.3% 378 of 30,024 new
Applebot Apple answer index 1.2% 356 of 30,024 new
Amzn-SearchBot Amazon answer index 0.8% 253 of 30,024 new
TikTokSpider ByteDance live retrieval 0.8% 251 of 30,024 new
Amzn-User Amazon live retrieval 0.7% 200 of 30,024 new
MistralAI-Index Mistral AI answer index 0.3% 102 of 30,024 new
facebookexternalhit Meta live retrieval 0.2% 54 of 30,024 new
Googlebot-News Google answer index 0.1% 37 of 30,024 new
Bingbot Microsoft answer index 0.1% 35 of 30,024 new
Meta-ExternalAds Meta training 0.1% 29 of 30,024 new
msnbot Microsoft answer index 0.1% 18 of 30,024 new
Googlebot Google answer index 0.1% 17 of 30,024 new
OAI-AdsBot OpenAI live retrieval 0.0% 12 of 30,024 new

Trend share of measured sites

% of sites20151050% of sites08-3009-2810-28block an answer engineblock a training crawlerpublish llms.txt

Where the blocking is share blocking an answer engine, by top-level domain

bar chart.uk (n=429)17%.fr (n=321)16.8%.de (n=819)16.1%.jp (n=412)9%.com (n=14674)8%.org (n=1833)6.9%.net (n=1297)6.8%.br (n=360)6.7%.in (n=301)3.3%.ru (n=1152)3%.io (n=656)2.7%.edu (n=472)2.3%

Only TLDs with at least 40 measured domains are shown, and each bar carries its sample size. A TLD is a weak proxy for jurisdiction. Full segmentation.

Changes recorded this week 0 total

No crawler policy changes were recorded in this week.

Full change log · RSS

Cite this CC BY 4.0

The figures on this page are frozen to the week shown and will not be edited after publication. Corrections are appended, not overwritten.

Crawl Census, "The state of AI crawling, 2002-W21", 2002-05-26. https://crawlcensus.com/report/2002-W21

Machine-readable: census totals, full corpus, crawler registry. Method: methodology.