Websites blocking cohere-training-data-crawler
Cohere's crawler for collecting web data used to train its enterprise language models.
Operated by Cohere · Training crawler
5.25%
block cohere-training-data-crawler
1.29 million of 24.68 million domains with a known robots.txt policy
1.84%
block it by name
455K domains, the rest inherit a blanket
* block
2.41%
restrict it by name
596K domains name it and close some of its paths
4.51%
name it in robots.txt
1.11 million domains give it rules of their own
Data for Oct 2026. All four percentages
divide by the same 24.68 million domains. Blocked counts both
explicit User-agent: cohere-training-data-crawler blocks and blanket
* blocks the crawler inherits.
Blocking trend
2%
4%
6%
Aug 2026
Sep 2026
Oct 2026
Share of domains with a known robots.txt policy
blocking cohere-training-data-crawler in each monthly crawl. The denominator moves with each crawl and is
defined in the methodology.
Historical mentions, HTTP Archive
1%
2%
3%
4%
Oct 2023
Jul 2024
Apr 2025
Jan 2026
Sep 2026
Share of sites in each month's
HTTP Archive crawl whose robots.txt named
cohere-training-data-crawler (or a legacy alias of it). This series counts
mentions only (the archive keeps no rule text), comes from a different population than our own
crawl above, and the two are never merged.
High-authority sites blocking cohere-training-data-crawler by name
| Domain | OPROpen Page Rank | Blocking since |
|---|---|---|
| cnn.com | 9.16/10 | Oct 2026 |
| amazon.co.jp | 9.16/10 | Sep 2026 |
| substack.com | 9.13/10 | Oct 2026 |
| amazon.de | 9.13/10 | Oct 2026 |
| usatoday.com | 9.08/10 | Sep 2026 |
| wired.com | 9.07/10 | Oct 2026 |
| upenn.edu | 9.00/10 | Oct 2026 |
| uol.com.br | 8.92/10 | Oct 2026 |
| amazon.ca | 8.91/10 | Oct 2026 |
| newyorker.com | 8.90/10 | Oct 2026 |
| mashable.com | 8.87/10 | Sep 2026 |
| francetvinfo.fr | 8.87/10 | Aug 2026 |
| arstechnica.com | 8.84/10 | Aug 2026 |
| unc.edu | 8.83/10 | Oct 2026 |
| radiofrance.fr | 8.82/10 | Aug 2026 |
| zeit.de | 8.81/10 | Sep 2026 |
| pcmag.com | 8.80/10 | Aug 2026 |
| welt.de | 8.78/10 | Oct 2026 |
| ardmediathek.de | 8.77/10 | Aug 2026 |
| fliphtml5.com | 8.75/10 | Oct 2026 |
| tagesschau.de | 8.73/10 | Aug 2026 |
| rfi.fr | 8.73/10 | Aug 2026 |
| verywellmind.com | 8.72/10 | Aug 2026 |
| bostonglobe.com | 8.71/10 | Oct 2026 |
| people.com | 8.70/10 | Aug 2026 |
| bild.de | 8.70/10 | Oct 2026 |
| ndr.de | 8.70/10 | Oct 2026 |
| mainichi.jp | 8.69/10 | Sep 2026 |
| amazon.com.br | 8.67/10 | Oct 2026 |
| vanityfair.com | 8.64/10 | Oct 2026 |
Sites whose robots.txt names cohere-training-data-crawler (or a legacy
alias) with a full block, ranked by Open Page Rank. Sites blocking every crawler with a blanket
* rule are counted above but not listed here. Since dates start at our first
observation of the rule.
Restricting some paths
- statista.com
- npr.org
- squarespace.com
- tripadvisor.com
- vox.com
- youradchoices.ca
- ccm19.de
- lifehacker.com
- nintendo.com
- asahi.com
- ign.com
- everydayhealth.com
Also matched via
the legacy token cohere-ai,
which some sites still use for this crawler.