Weekly report · 2003-W11

The state of AI crawling

Measured on 30,590 reachable domains for the week ending 2003-03-16. 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.

← 2003-W10  ·  2003-W12 →

30,590
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 7.4% 2,262 of 30,589 new
CCBot Common Crawl Foundation archive 7.1% 2,160 of 30,589 new
Bytespider ByteDance training 6.8% 2,078 of 30,589 new
ClaudeBot Anthropic training 6.1% 1,860 of 30,589 new
Amazonbot Amazon training 5.1% 1,575 of 30,589 new
Applebot-Extended Apple training 5.1% 1,558 of 30,589 new
Meta-ExternalAgent Meta training 5.1% 1,556 of 30,589 new
Google-Extended Google training 5.1% 1,548 of 30,589 new
anthropic-ai Anthropic training 4.8% 1,453 of 30,589 new
omgilibot Webz.io training 4.7% 1,444 of 30,589 new
Diffbot Diffbot training 4.5% 1,378 of 30,589 new
cohere-ai Cohere live retrieval 4.5% 1,364 of 30,589 new
omgili Webz.io training 4.4% 1,331 of 30,589 new
SemrushBot-OCOB Semrush training 3.6% 1,096 of 30,589 new
PerplexityBot Perplexity answer index 3.4% 1,029 of 30,589 new
ChatGPT-User OpenAI live retrieval 3.3% 1,018 of 30,589 new
FacebookBot Meta training 3.1% 943 of 30,589 new
YouBot You.com answer index 2.9% 892 of 30,589 new
ImagesiftBot ImageSift (Hive) training 2.7% 829 of 30,589 new
Timpibot Timpi training 2.7% 818 of 30,589 new
Scrapy Zyte (open-source framework) training 2.4% 724 of 30,589 new
AI2Bot Allen Institute for AI training 2.3% 703 of 30,589 new
OAI-SearchBot OpenAI answer index 2.0% 627 of 30,589 new
cohere-training-data-crawler Cohere training 2.0% 616 of 30,589 new
Google-CloudVertexBot Google live retrieval 1.9% 590 of 30,589 new
GoogleOther Google training 1.8% 554 of 30,589 new
Meta-ExternalFetcher Meta live retrieval 1.7% 526 of 30,589 new
DuckAssistBot DuckDuckGo live retrieval 1.7% 522 of 30,589 new
PanguBot Huawei training 1.7% 516 of 30,589 new
Claude-User Anthropic live retrieval 1.6% 503 of 30,589 new
Claude-SearchBot Anthropic answer index 1.6% 480 of 30,589 new
Perplexity-User Perplexity live retrieval 1.6% 478 of 30,589 new
MistralAI-User Mistral AI live retrieval 1.5% 471 of 30,589 new
Kangaroo Bot Kangaroo LLM training 1.5% 452 of 30,589 new
Meta-WebIndexer Meta answer index 1.2% 361 of 30,589 new
Applebot Apple answer index 1.2% 358 of 30,589 new
MistralAI-Training Mistral AI training 1.1% 335 of 30,589 new
TikTokSpider ByteDance live retrieval 0.9% 261 of 30,589 new
Amzn-SearchBot Amazon answer index 0.8% 249 of 30,589 new
Amzn-User Amazon live retrieval 0.6% 187 of 30,589 new
MistralAI-Index Mistral AI answer index 0.3% 96 of 30,589 new
facebookexternalhit Meta live retrieval 0.2% 48 of 30,589 new
Googlebot-News Google answer index 0.1% 36 of 30,589 new
Bingbot Microsoft answer index 0.1% 33 of 30,589 new
Meta-ExternalAds Meta training 0.1% 24 of 30,589 new
msnbot Microsoft answer index 0.1% 18 of 30,589 new
Googlebot Google answer index 0.1% 17 of 30,589 new
OAI-AdsBot OpenAI live retrieval 0.0% 13 of 30,589 new

Trend share of measured sites

% of sites20151050% of sites08-2409-2210-22block 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=448)16.5%.fr (n=331)16.3%.de (n=838)15.6%.jp (n=406)8.9%.br (n=369)7.3%.com (n=14942)7.1%.org (n=1849)5.4%.net (n=1313)5.1%.in (n=311)3.2%.ru (n=1154)2.9%.edu (n=470)2.3%.io (n=657)2.1%

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, 2003-W11", 2003-03-16. https://crawlcensus.com/report/2003-W11

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