Weekly report · 2016-W12

The state of AI crawling

Measured on 30,487 reachable domains for the week ending 2016-03-27. 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.

← 2016-W11  ·  2016-W13 →

30,487
domains measured
6%
block an answer engine
19%
refuse AI agents at the edge
13%
publish an llms.txt

Block rate by crawler with change on the previous week

CrawlerOperatorContent used forBlock rateSitesWeek on week
GPTBot OpenAI training 11.2% 3,427 of 30,484 new
CCBot Common Crawl Foundation archive 11.1% 3,380 of 30,484 new
Bytespider ByteDance training 10.7% 3,247 of 30,484 new
ClaudeBot Anthropic training 9.9% 3,013 of 30,484 new
Meta-ExternalAgent Meta training 9.0% 2,734 of 30,484 new
Amazonbot Amazon training 8.9% 2,719 of 30,484 new
Google-Extended Google training 8.9% 2,714 of 30,484 new
Applebot-Extended Apple training 8.7% 2,661 of 30,484 new
anthropic-ai Anthropic training 4.1% 1,259 of 30,484 new
omgilibot Webz.io training 4.1% 1,242 of 30,484 new
cohere-ai Cohere live retrieval 3.8% 1,153 of 30,484 new
Diffbot Diffbot training 3.8% 1,150 of 30,484 new
SemrushBot-OCOB Semrush training 3.7% 1,133 of 30,484 new
omgili Webz.io training 3.7% 1,127 of 30,484 new
PerplexityBot Perplexity answer index 3.5% 1,082 of 30,484 new
ChatGPT-User OpenAI live retrieval 3.5% 1,074 of 30,484 new
FacebookBot Meta training 3.2% 968 of 30,484 new
YouBot You.com answer index 2.9% 883 of 30,484 new
ImagesiftBot ImageSift (Hive) training 2.7% 830 of 30,484 new
Timpibot Timpi training 2.7% 826 of 30,484 new
Scrapy Zyte (open-source framework) training 2.4% 732 of 30,484 new
OAI-SearchBot OpenAI answer index 2.2% 676 of 30,484 new
AI2Bot Allen Institute for AI training 2.2% 674 of 30,484 new
cohere-training-data-crawler Cohere training 2.0% 595 of 30,484 new
Meta-ExternalFetcher Meta live retrieval 1.7% 531 of 30,484 new
DuckAssistBot DuckDuckGo live retrieval 1.7% 523 of 30,484 new
PanguBot Huawei training 1.7% 503 of 30,484 new
Claude-SearchBot Anthropic answer index 1.6% 476 of 30,484 new
Claude-User Anthropic live retrieval 1.5% 468 of 30,484 new
Kangaroo Bot Kangaroo LLM training 1.5% 450 of 30,484 new
Perplexity-User Perplexity live retrieval 1.4% 440 of 30,484 new
MistralAI-User Mistral AI live retrieval 1.4% 421 of 30,484 new
Google-CloudVertexBot Google live retrieval 1.2% 355 of 30,484 new
Meta-WebIndexer Meta answer index 1.2% 352 of 30,484 new
Applebot Apple answer index 1.1% 339 of 30,484 new
GoogleOther Google training 1.1% 332 of 30,484 new
TikTokSpider ByteDance live retrieval 0.8% 255 of 30,484 new
Amzn-SearchBot Amazon answer index 0.7% 213 of 30,484 new
Amzn-User Amazon live retrieval 0.4% 130 of 30,484 new
MistralAI-Index Mistral AI answer index 0.2% 73 of 30,484 new
facebookexternalhit Meta live retrieval 0.2% 59 of 30,484 new
MistralAI-Training Mistral AI training 0.2% 58 of 30,484 new
Bingbot Microsoft answer index 0.1% 40 of 30,484 new
Googlebot-News Google answer index 0.1% 39 of 30,484 new
msnbot Microsoft answer index 0.1% 27 of 30,484 new
Meta-ExternalAds Meta training 0.1% 20 of 30,484 new
Googlebot Google answer index 0.1% 16 of 30,484 new
OAI-AdsBot OpenAI live retrieval 0.0% 3 of 30,484 new

Trend share of measured sites

% of sites20151050% of sites08-2209-0309-15block 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)17.2%.fr (n=327)16.5%.de (n=838)16.3%.jp (n=417)9.1%.com (n=14955)6.5%.br (n=380)5.5%.org (n=1686)5.3%.net (n=1322)3.7%.ru (n=1168)3.1%.in (n=302)2.6%.edu (n=476)2.3%.io (n=658)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, 2016-W12", 2016-03-27. https://crawlcensus.com/report/2016-W12

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