ProminaraBotvisited because you asked it to.
Prominara reads a website only when the account that owns it runs a scan. The scan measures how well each page can be cited by AI answer engines. It never crawls sites nobody registered, and it stops at your plan's page limit.
Requests first look like an ordinary browser. When a firewall challenges those, one final request identifies itself with the user agent below so you can allow exactly that, without loosening bot protection for anyone else.
Mozilla/5.0 (compatible; ProminaraBot/1.0; +https://prominara.com/crawler)- User agent token
- ProminaraBot
- Trigger
- A site owner runs a scan in Prominara
- Pages per scan
- Up to the plan limit, never the whole site
- robots.txt
- Honored
- Content use
- Analysis for that account only. Never trained on, never resold.
Allow the crawler in about two minutes.
Your AI visibility tracking keeps working either way: prompt checks ask the AI engines directly. The scan is what powers the page audit, the missions and change detection, so it is worth two minutes.
Cloudflare
Security → WAF → Custom rules → Create rule. Expression: User Agent contains "ProminaraBot". Action: Skip → all remaining custom rules, and also skip Bot Fight Mode / Super Bot Fight Mode. Deploy, then rerun the scan in Prominara.
Vercel
Project → Firewall → Add rule. Condition: User Agent contains "ProminaraBot". Action: Bypass. Save, then rerun the scan.
Akamai, DataDome, PerimeterX and others
Add "ProminaraBot" to the bot allowlist or create a bypass rule keyed on the user agent. If your vendor requires an IP allowlist instead, contact us: the crawler runs from Vercel's US-East (AWS) egress and its addresses rotate, so a user-agent rule is the reliable option.
Questions, or want us to stop visiting a site? Contact us