Scanning policy
Kipra runs internet-wide research scans to measure how much AI infrastructure is exposed to the internet. This data powers the AI Internet Platform and our public research. If you saw traffic from us, this page explains it.
What our scanners do
- Connect to publicly reachable services and read what they show anyone: banners, version endpoints and public metadata.
- Identify the product (for example, a model server or vector database) and whether it requires authentication.
What our scanners never do
- Log in, guess passwords or use credentials of any kind.
- Read, download or modify data.
- Exploit vulnerabilities or attempt to disrupt services.
How to recognize us
| Item | Value |
|---|---|
| User agent | KipraResearch/1.0 (+https://kipra.dev/scanning) |
| Source IPs | Published on this page before scanning begins. |
Opt out
Email optout@kipra.dev with the IP addresses or ranges you'd like excluded. We honor requests within 5 working days and remove past results for those addresses from our data.
Found something of yours?
If our research finds an exposed system we can attribute to your organization, we'll contact you privately as described in our disclosure policy.