Weekly report · 2026-W36
The state of AI crawling
Measured on 30,224 reachable domains for the week ending 2026-09-06. 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.
30,224
domains measured
7%
block an answer engine
19%
refuse AI agents at the edge
14%
publish an llms.txt
Block rate by crawler with change on the previous week
| Crawler | Operator | Content used for | Block rate | Sites | Week on week |
|---|---|---|---|---|---|
| GPTBot | OpenAI | training | 11.4% | 3,492 of 30,566 | +0.1 pt |
| CCBot | Common Crawl Foundation | archive | 11.2% | 3,429 of 30,566 | +0.2 pt |
| Bytespider | ByteDance | training | 10.8% | 3,293 of 30,566 | +0.2 pt |
| ClaudeBot | Anthropic | training | 10.0% | 3,063 of 30,566 | +0.1 pt |
| Google-Extended | training | 9.0% | 2,764 of 30,566 | +0.1 pt | |
| Meta-ExternalAgent | Meta | training | 9.0% | 2,760 of 30,566 | +0.1 pt |
| Amazonbot | Amazon | training | 9.0% | 2,748 of 30,566 | +0.1 pt |
| Applebot-Extended | Apple | training | 8.8% | 2,684 of 30,566 | +0.1 pt |
| anthropic-ai | Anthropic | training | 4.2% | 1,290 of 30,566 | +0.1 pt |
| omgilibot | Webz.io | training | 4.2% | 1,279 of 30,566 | +0.1 pt |
| Diffbot | Diffbot | training | 3.9% | 1,181 of 30,566 | +0.1 pt |
| omgili | Webz.io | training | 3.8% | 1,161 of 30,566 | +0.1 pt |
| cohere-ai | Cohere | live retrieval | 3.8% | 1,152 of 30,566 | +0.1 pt |
| SemrushBot-OCOB | Semrush | training | 3.8% | 1,147 of 30,566 | +0.1 pt |
| PerplexityBot | Perplexity | answer index | 3.6% | 1,115 of 30,566 | +0.1 pt |
| ChatGPT-User | OpenAI | live retrieval | 3.6% | 1,112 of 30,566 | +0.1 pt |
| FacebookBot | Meta | training | 3.3% | 994 of 30,566 | +0.1 pt |
| YouBot | You.com | answer index | 2.8% | 868 of 30,566 | +0.1 pt |
| ImagesiftBot | ImageSift (Hive) | training | 2.8% | 843 of 30,566 | +0.1 pt |
| Timpibot | Timpi | training | 2.7% | 835 of 30,566 | +0.1 pt |
| Scrapy | Zyte (open-source framework) | training | 2.3% | 716 of 30,566 | +0.1 pt |
| OAI-SearchBot | OpenAI | answer index | 2.3% | 694 of 30,566 | +0.1 pt |
| AI2Bot | Allen Institute for AI | training | 2.2% | 685 of 30,566 | +0.1 pt |
| cohere-training-data-crawler | Cohere | training | 1.9% | 595 of 30,566 | level |
| DuckAssistBot | DuckDuckGo | live retrieval | 1.7% | 532 of 30,566 | level |
| Meta-ExternalFetcher | Meta | live retrieval | 1.7% | 530 of 30,566 | +0.1 pt |
| PanguBot | Huawei | training | 1.7% | 517 of 30,566 | level |
| Claude-SearchBot | Anthropic | answer index | 1.6% | 482 of 30,566 | level |
| Claude-User | Anthropic | live retrieval | 1.6% | 474 of 30,566 | +0.1 pt |
| Kangaroo Bot | Kangaroo LLM | training | 1.5% | 459 of 30,566 | level |
| Perplexity-User | Perplexity | live retrieval | 1.5% | 450 of 30,566 | level |
| MistralAI-User | Mistral AI | live retrieval | 1.4% | 428 of 30,566 | +0.1 pt |
| Applebot | Apple | answer index | 1.2% | 365 of 30,566 | level |
| Google-CloudVertexBot | live retrieval | 1.2% | 355 of 30,566 | level | |
| Meta-WebIndexer | Meta | answer index | 1.1% | 349 of 30,566 | +0.1 pt |
| GoogleOther | training | 1.1% | 337 of 30,566 | +0.1 pt | |
| TikTokSpider | ByteDance | live retrieval | 0.8% | 254 of 30,566 | level |
| Amzn-SearchBot | Amazon | answer index | 0.7% | 217 of 30,566 | level |
| Amzn-User | Amazon | live retrieval | 0.4% | 129 of 30,566 | level |
| MistralAI-Index | Mistral AI | answer index | 0.2% | 67 of 30,566 | level |
| facebookexternalhit | Meta | live retrieval | 0.2% | 56 of 30,566 | level |
| MistralAI-Training | Mistral AI | training | 0.2% | 55 of 30,566 | level |
| Bingbot | Microsoft | answer index | 0.1% | 41 of 30,566 | level |
| Googlebot-News | answer index | 0.1% | 39 of 30,566 | level | |
| msnbot | Microsoft | answer index | 0.1% | 25 of 30,566 | level |
| Meta-ExternalAds | Meta | training | 0.1% | 20 of 30,566 | level |
| Googlebot | answer index | 0.1% | 16 of 30,566 | level | |
| OAI-AdsBot | OpenAI | live retrieval | 0.0% | 2 of 30,566 | level |
Trend share of measured sites
Where the blocking is share blocking an answer engine, by top-level domain
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 200 total
- bayt.com AI access score moved 66 to 54
- geekbench.com GPTBot is now allowed by robots.txt
- geekbench.com ClaudeBot is now allowed by robots.txt
- geekbench.com Google-Extended is now allowed by robots.txt
- geekbench.com Applebot-Extended is now allowed by robots.txt
- geekbench.com Meta-ExternalAgent is now allowed by robots.txt
- geekbench.com Bytespider is now allowed by robots.txt
- geekbench.com Amazonbot is now allowed by robots.txt
- geekbench.com CCBot is now allowed by robots.txt
- openpr.com OAI-SearchBot is no longer refused at the network edge
- topps.com AI access score moved 61 to 49
- btcscan.org GPTBot is no longer refused at the network edge
- btcscan.org OAI-SearchBot is no longer refused at the network edge
- btcscan.org ClaudeBot is no longer refused at the network edge
- btcscan.org PerplexityBot is no longer refused at the network edge
- btcscan.org AI access score moved 33 to 46
- neh.gov AI access score moved 54 to 82
- ulg.ac.be ClaudeBot is no longer refused at the network edge
- silive.com YouBot is now blocked by robots.txt
- silive.com Scrapy is now blocked by robots.txt
- pngegg.com SemrushBot-OCOB is now allowed by robots.txt
- revues.org GPTBot is no longer refused at the network edge
- revues.org ClaudeBot is no longer refused at the network edge
- hbvl.be GPTBot is now allowed by robots.txt
- hbvl.be OAI-SearchBot is now allowed by robots.txt
- hbvl.be ChatGPT-User is now allowed by robots.txt
- hbvl.be OAI-AdsBot is now allowed by robots.txt
- hbvl.be ClaudeBot is now allowed by robots.txt
- hbvl.be Claude-User is now allowed by robots.txt
- hbvl.be Claude-SearchBot is now allowed by robots.txt
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, 2026-W36", 2026-09-06. https://crawlcensus.com/report/2026-W36Machine-readable: census totals, full corpus, crawler registry. Method: methodology.