ElevenlabsCrawler
ElevenLabs' crawler for content that may train its generative AI models. On our sites it sent more requests than any other crawler, almost all in one day.
| Operator | ElevenLabs |
|---|---|
| Type | AI: AI training |
| robots.txt token | ElevenlabsCrawler |
| Follows robots.txt | Yes, by the operator's account; Crawl-delay not documented |
| Sends visitors | No |
| Links to your pages | No |
| Can end up in AI training | Yes |
| Verification |
None published
Our tracker does not verify it yet. |
| Our call |
Publisher: Block E-commerce: Block Expert site: No strong call What the profiles mean |
How to recognize it
User agent, as it reaches our logs (identical to ElevenLabs’ docs):
Mozilla/5.0 (compatible; ElevenlabsCrawler/1.0; +https://elevenlabs.io/docs/bots)
Match on ElevenlabsCrawler. Its sibling ElevenlabsBot/1.0 is a different agent with its own token.
- robots.txt token:
ElevenlabsCrawler - IP list: none published.
- Reverse DNS: none published.
What it does
ElevenLabs says ElevenlabsCrawler collects web content that may be used to train its generative AI models, and that disallowing it tells ElevenLabs not to use the site’s content for training. Changes to robots.txt take effect within about 24 hours. ElevenlabsBot is separate: it fetches the pages that ElevenLabs customers add to their voice agents’ knowledge bases.
What you gain by allowing it
- Little that a website can point to. ElevenLabs’ models generate speech and voice agents; there is no search result, link or citation in return.
What you lose
By allowing it
- Content used for training with nothing in return.
- Load in bursts. On our sites it sent 8,347 requests in a single day.
By blocking it
- Nothing visible, as long as you block only
ElevenlabsCrawler. Customers’ voice agents fetch throughElevenlabsBot, which a separate group controls. An expert site whose content customers want in their agents should keep that one open.
How to verify it
ElevenLabs publishes no IP list or reverse DNS pattern, and Cloudflare does not flag it as a verified bot, so there is no way to prove a request is real. The user agent matched ElevenLabs’ documented string exactly, but the requests came from about ten different networks, mostly hosting providers, which is also what a scraper hiding behind a borrowed name looks like.
Gotchas and history
- Two bots, one company: a rule for
ElevenlabsBotdoes not coverElevenlabsCrawler, and the reverse. ElevenLabs states that each setting is independent. - No verification path: with no list and no Cloudflare flag, blocking by user agent is all a firewall can do. robots.txt is the documented route.
What we observed
Data: Cloudflare Logpush for two of our sites, 19-26 October 2026 (8 days).
- 8,675 requests carried this user agent, the most of any crawler in the period; ClaudeBot had 4,937.
- 8,347 of them came on 22 October, and 288 on the day before. After that only a few dozen arrived.
- They covered 5,472 different paths, nearly all on the static content site, and 8,665 got a 200. 15 were for
/robots.txt. - The requests came from ten networks, among them Vultr, M247, ColoCrossing, Latitude.sh and one US consumer ISP. None was flagged by Cloudflare. Whether they were ElevenLabs we cannot tell.
About ElevenLabs
- ElevenlabsCrawler collects content that may train ElevenLabs’ generative AI models.
- ElevenlabsBot fetches pages that customers add to their voice agents’ knowledge bases, and refreshes them.
ElevenLabs says each robots.txt setting is independent and changes apply within about 24 hours. It publishes no IP list.