How to allow, verify or block Perplexity (Comet and Perplexity-User)
Allow the Perplexity-User token and its published IP ranges at your firewall. Perplexity also asks sites to allow PerplexityBot, its search crawler, in robots.txt.
Checked October 20, 2026. Comet runs on the user's device with their sessions and is reported to look like ordinary Chromium, so it cannot be reliably identified.
How it identifies itself
| Signal | Value | Documented by the maker |
|---|---|---|
| user-agent | Perplexity-User/1.0 (+https://perplexity.ai/perplexity-user) | Yes docs.perplexity.ai |
| ip-list | https://www.perplexity.com/perplexity-user.json | Yes docs.perplexity.ai |
| user-agent | No Comet-specific identifier found in Perplexity's bot docs | No, third-party only docs.perplexity.ai |
- Allow it
- Allow the Perplexity-User token and its published IP ranges at your firewall. Perplexity also asks sites to allow PerplexityBot, its search crawler, in robots.txt.
- Verify it
- Check that requests claiming Perplexity-User come from the ranges at perplexity.com/perplexity-user.json. There is no documented way to single out the Comet browser agent.
- Block it
- Block the Perplexity-User token or IP ranges at the firewall. Perplexity says this user-triggered fetcher generally ignores robots.txt, so a robots.txt rule will not stop it.
robots.txt lines for Perplexity (Comet and Perplexity-User)
Copy the group you want for Perplexity-User. The most specific matching group wins over your * group.
Let it in
User-agent: Perplexity-User
Allow: /Keep it out
User-agent: Perplexity-User
Disallow: /What we can and can't confirm
1 of the 3 signals above come from third parties, not from the maker, so treat them as hints. Matching a user agent string alone is easy to fake. A published IP list or a signed request is the stronger check, where the maker offers one.