Weekly report · 2005-W38

The state of AI crawling

Measured on 30,250 reachable domains for the week ending 2005-09-25. 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.

← 2005-W37  ·  2005-W39 →

30,250
domains measured
7%
block an answer engine
20%
refuse AI agents at the edge
14%
publish an llms.txt

Block rate by crawler with change on the previous week

CrawlerOperatorContent used forBlock rateSitesWeek on week
GPTBot OpenAI training 7.1% 2,152 of 30,238 new
CCBot Common Crawl Foundation archive 6.7% 2,039 of 30,238 new
Bytespider ByteDance training 6.5% 1,964 of 30,238 new
ClaudeBot Anthropic training 5.8% 1,749 of 30,238 new
Amazonbot Amazon training 4.8% 1,440 of 30,238 new
Meta-ExternalAgent Meta training 4.7% 1,429 of 30,238 new
Applebot-Extended Apple training 4.7% 1,422 of 30,238 new
Google-Extended Google training 4.7% 1,413 of 30,238 new
anthropic-ai Anthropic training 4.4% 1,328 of 30,238 new
omgilibot Webz.io training 4.4% 1,328 of 30,238 new
Diffbot Diffbot training 4.2% 1,257 of 30,238 new
cohere-ai Cohere live retrieval 4.1% 1,246 of 30,238 new
omgili Webz.io training 4.0% 1,217 of 30,238 new
SemrushBot-OCOB Semrush training 3.7% 1,123 of 30,238 new
ChatGPT-User OpenAI live retrieval 3.4% 1,037 of 30,238 new
PerplexityBot Perplexity answer index 3.4% 1,029 of 30,238 new
FacebookBot Meta training 3.2% 965 of 30,238 new
YouBot You.com answer index 3.0% 895 of 30,238 new
ImagesiftBot ImageSift (Hive) training 2.8% 840 of 30,238 new
Timpibot Timpi training 2.7% 827 of 30,238 new
Scrapy Zyte (open-source framework) training 2.4% 726 of 30,238 new
AI2Bot Allen Institute for AI training 2.3% 699 of 30,238 new
OAI-SearchBot OpenAI answer index 2.1% 624 of 30,238 new
cohere-training-data-crawler Cohere training 2.0% 613 of 30,238 new
Meta-ExternalFetcher Meta live retrieval 1.7% 525 of 30,238 new
DuckAssistBot DuckDuckGo live retrieval 1.7% 520 of 30,238 new
PanguBot Huawei training 1.7% 515 of 30,238 new
Claude-User Anthropic live retrieval 1.6% 493 of 30,238 new
Claude-SearchBot Anthropic answer index 1.6% 471 of 30,238 new
Perplexity-User Perplexity live retrieval 1.5% 467 of 30,238 new
MistralAI-User Mistral AI live retrieval 1.5% 457 of 30,238 new
Google-CloudVertexBot Google live retrieval 1.5% 456 of 30,238 new
Kangaroo Bot Kangaroo LLM training 1.5% 452 of 30,238 new
GoogleOther Google training 1.4% 418 of 30,238 new
Meta-WebIndexer Meta answer index 1.2% 357 of 30,238 new
Applebot Apple answer index 1.2% 350 of 30,238 new
TikTokSpider ByteDance live retrieval 0.9% 258 of 30,238 new
Amzn-SearchBot Amazon answer index 0.8% 238 of 30,238 new
MistralAI-Training Mistral AI training 0.6% 189 of 30,238 new
Amzn-User Amazon live retrieval 0.6% 177 of 30,238 new
MistralAI-Index Mistral AI answer index 0.3% 91 of 30,238 new
facebookexternalhit Meta live retrieval 0.2% 54 of 30,238 new
Bingbot Microsoft answer index 0.1% 41 of 30,238 new
Googlebot-News Google answer index 0.1% 35 of 30,238 new
Meta-ExternalAds Meta training 0.1% 23 of 30,238 new
msnbot Microsoft answer index 0.1% 19 of 30,238 new
Googlebot Google answer index 0.0% 15 of 30,238 new
OAI-AdsBot OpenAI live retrieval 0.0% 12 of 30,238 new

Trend share of measured sites

% of sites20151050% of sites08-2209-1910-17block 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=441)16.3%.de (n=833)16%.fr (n=328)14.9%.jp (n=418)9.1%.com (n=14818)6.9%.br (n=377)5.6%.org (n=1661)5.5%.net (n=1309)4.3%.in (n=303)3.3%.ru (n=1156)2.9%.edu (n=467)2.4%.io (n=654)2%

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, 2005-W38", 2005-09-25. https://crawlcensus.com/report/2005-W38

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