AI-Access reference · read 26 Oct 2026 · first reading 23 Oct 2026
EU AI-Access reference: how AI crawlers are treated by country and CMS
On 26 Oct 2026, 18.5% of 10,355 popular EU websites blocked at least one of GPTBot, ClaudeBot, Google-Extended and CCBot in robots.txt: from 11.1% of .si sites (Slovenia) to 31.0% of .de sites (Germany). 11.7% publish an llms.txt.
By country-code domain
| Country | Sites read | Block GPTBot | Block at least one of the four training crawlers | Block an answer-engine search crawler | Publish llms.txt |
|---|---|---|---|---|---|
| Germany (.de) | 462 | 27.3% | 31.0% | 16.2% | 8.8% |
| Italy (.it) | 427 | 21.8% | 27.9% | 15.2% | 9.3% |
| Poland (.pl) | 426 | 23.9% | 26.3% | 6.8% | 10.2% |
| Finland (.fi) | 396 | 22.7% | 25.3% | 13.9% | 10.5% |
| France (.fr) | 437 | 21.1% | 24.5% | 13.3% | 4.9% |
| Netherlands (.nl) | 427 | 22.2% | 23.4% | 15.7% | 12.5% |
| Spain (.es) | 425 | 18.6% | 22.6% | 11.8% | 9.8% |
| Denmark (.dk) | 431 | 20.4% | 22.5% | 9.7% | 15.9% |
| Sweden (.se) | 455 | 17.8% | 19.8% | 11.6% | 10.1% |
| Czechia (.cz) | 458 | 12.9% | 17.2% | 6.6% | 9.9% |
| Belgium (.be) | 416 | 14.2% | 16.6% | 8.7% | 11.1% |
| Austria (.at) | 411 | 13.6% | 16.5% | 6.3% | 7.5% |
| Lithuania (.lt) | 408 | 13.0% | 16.4% | 4.9% | 16.0% |
| Slovakia (.sk) | 422 | 11.1% | 16.4% | 4.3% | 11.1% |
| Ireland (.ie) | 424 | 13.9% | 16.3% | 6.8% | 13.8% |
| Portugal (.pt) | 414 | 13.5% | 16.2% | 3.6% | 12.7% |
| Latvia (.lv) | 390 | 13.3% | 15.1% | 4.4% | 11.5% |
| Greece (.gr) | 418 | 10.5% | 15.1% | 3.1% | 14.0% |
| Luxembourg (.lu) | 153 | 14.4% | 15.0% | 8.5% | 12.8% |
| Estonia (.ee) | 439 | 10.5% | 14.1% | 3.9% | 17.6% |
| Croatia (.hr) | 423 | 10.4% | 13.0% | 3.3% | 12.1% |
| Romania (.ro) | 427 | 8.9% | 12.6% | 2.3% | 21.0% |
| Bulgaria (.bg) | 426 | 8.5% | 12.2% | 3.3% | 10.4% |
| Hungary (.hu) | 435 | 10.6% | 12.2% | 3.9% | 7.1% |
| Slovenia (.si) | 405 | 9.1% | 11.1% | 3.5% | 13.9% |
Sorted by the fourth column; each country has its own page with the per-crawler table.
By CMS
| CMS (home-page signature) | Sites read | Block GPTBot | Block at least one of the four training crawlers | Block an answer-engine search crawler | Publish llms.txt |
|---|---|---|---|---|---|
| Next.js | 583 | 18.7% | 20.8% | 10.5% | 15.7% |
| PrestaShop | 139 | 10.1% | 18.0% | 2.9% | 7.1% |
| Magento | 205 | 8.8% | 11.2% | 2.0% | 21.0% |
| WordPress | 1,655 | 9.2% | 10.4% | 4.4% | 15.2% |
| Drupal | 528 | 8.1% | 9.3% | 4.5% | 7.7% |
| TYPO3 | 132 | 6.8% | 7.6% | 3.0% | 7.8% |
CMSs with at least 100 sites read. Between 23 Oct 2026 and 26 Oct 2026, 64 of the 10,355 sites read in both runs changed what they ask at least one of the 14 AI crawlers to do (46 started blocking, 18 stopped). Run 1 and run history: EU AI-Access Index, EU llms.txt Index.
Method and limits
- Sample: the top 600 domains by Tranco rank under each of 25 EU country-code domains (Luxembourg 250), Tranco list 56Z8N. The same sites are read again each week, so the dates on this page are repeat readings of one list, not new samples.
- What is read:
/robots.txtof each site, per RFC 9309; "blocks" means the crawler may not fetch/. Shares use the sites whose robots.txt was read as base; the others are coverage, not "allows". llms.txt: the file at/llms.txtis valid if it passes the llmstxt.org shape check (method). - Limits: robots.txt states a request, not what a firewall enforces, and this index does not measure what any assistant cites. A country-code domain is not the operator's country, and a CMS is a signature in the home page, not a verified platform. Popular sites only: a random sample of 24,550 .de, .at and .ch hosts reads lower, 2.27% for GPTBot (Ahrefs, Aug 2026).
- Data: One row per domain, 23 Oct 2026 · One row per domain, 26 Oct 2026 · llms.txt per domain, 25 Oct 2026. Licence CC BY 4.0; cite as "CitedBy EU AI-Access Index, 26 Oct 2026".
What does your robots.txt say to AI crawlers?
The free check shows which AI crawlers your robots.txt and your server let in, with the exact fix. With the paid audit, we crawl your site up to your plan's page limit and give you the exact fix for each issue. From €9, 14-day money-back guarantee.
Checked on 26 Oct 2026 · Method: How CitedBy checks · Change log: What changed