AI IP Check FAQ
The most common questions about using this site's checks. Answers are shared with the tool pages they reference, and each entry links to the page where you can act on it.
Basics: purity, IP types, and dual stack
What does IP purity mean, and why do AI services care?
Purity is a community shorthand for a set of exit attributes: residential or data-center, native or broadcast, whether the address is flagged as proxy, VPN, or Tor, and whether it has an abuse history. This site does not publish a single purity score. It lists those dimensions separately, because different AI services weight them differently and one combined number would hide that.
What is the difference between a data-center IP and a residential IP?
A residential IP is assigned by a home broadband provider to an individual line. A data-center IP belongs to a hosting provider, where one block may be shared by many users and is more likely to be treated as proxy or automated traffic. The difference is not speed but how risk systems read the exit.
What is the difference between a native IP and a broadcast IP?
A native IP is registered and used in the same region. A broadcast IP is a block from another region announced locally, so its registration and its actual location differ. AI services usually judge region by registration, which is why a broadcast IP can appear to be somewhere you are not.
Why should IPv4 and IPv6 be checked separately?
A dual-stack network can use different exits, ASNs, regions, or proxy paths for IPv4 and IPv6. Checking only one address family can miss the path an AI service actually uses.
What an AI IP check is and how to use it
What is an AI IP check?
An AI IP check is network and availability testing aimed at AI services: on top of the IP profile (exit, ASN, region, network attributes), it adds connectivity to AI services' public resources, comparison with official supported regions, and routing-exit verification. It answers whether the current network can reach an AI service normally and where an issue sits. See the full definition article.
How is a manual IP lookup different from the automatic check?
The automatic check uses the actual network path of your current browser or command-line session and verifies IPv4 and IPv6 connectivity to public resources separately. A manually entered IP returns that address's static profile and regional reference, not a live test of your connection. For troubleshooting 'why the service fails right now', use the automatic check.
Can I run the check from a server command line?
Yes. The read-only script is at /sh, with full usage and result guidance at /cli. Review the script before running it. Its output describes the network, not account status.
Why do some AI services not show an exit IP?
Only services that reliably return a verifiable public exit IP are shown. A service without a suitable endpoint, an unavailable endpoint, or an untrustworthy response does not receive an exit conclusion.
Is the DNS exit the same as my public IP?
Not necessarily. A public IP comes from this site's HTTP request, while a DNS exit comes from a recursive DNS request reaching the controlled authoritative DNS server. Network policies can make the paths differ.
Reading results: verification, regions, and exits
What does it mean when ChatGPT reports a VPN or suspicious activity?
That message comes from the service's own judgement about the connection, usually related to the exit IP type, region, or reputation. Use this page to review the current exit region, ASN, whether it is flagged as data-center or proxy, and whether the IPv4 and IPv6 paths agree. The message alone is not an account restriction.
Why does ChatGPT keep asking for human verification?
Repeated verification and a page that will not load are different problems. Verification is more often tied to exit reputation, region consistency, or browser session state, while a failure to load starts with the public entry and region coverage. This page separates the two kinds of evidence.
How do I tell whether this exit suits Claude?
This page does not give a single suitability score. It lists the signals you can verify separately: exit region, ASN, residential or data-center type, proxy and Tor flags, and whether IPv4 and IPv6 agree. Those signals explain access differences; they do not determine whether an account is restricted.
What happens if my IPv4 and IPv6 exits differ on Gemini?
Gemini judges availability by the exit region it actually sees. If IPv4 and IPv6 land in different regions you can get a service that works sometimes and not others. This page shows the region and ASN for each address family so you can see which path triggers the limit.
Is a Gemini region error always a network problem?
No. A region message can come from the exit region, from the account's own region, or from a Workspace policy. Confirm whether the current exit region is supported first, then check the account and organization settings.
Is a different IPv4 and IPv6 DNS exit normal?
It can be. A dual-stack network can use different DNS, proxy, or routing policies for IPv4 and IPv6. This check displays them separately and compares each only with the matching site exit.
Boundaries: what results cannot prove
Does a data-center, proxy, or risk signal mean an account will be restricted?
No. These labels describe observable properties of the current network. Account outcomes can also depend on sign-in, behavior, subscription, organization policy, regional eligibility, and service-side rules.
The ChatGPT website opens but the app does not — is that an IP problem?
The website and the client may use different domains and resolution paths, and the client may also have cached state or a version issue. Check the website first: if only the app fails, the cause is more likely client-side; if both fail, review exit region and public-entry connectivity.
Does one shared exit mean routing is not active?
Not necessarily. Multiple AI domains can be assigned to the same proxy exit, or they may all connect directly. Compare the result with your rules and current public IP.
Does a different DNS exit prove a DNS leak?
No. It only shows a difference in this check's paths. Review secure DNS, proxy coverage, dual-stack rules, and the actual network setup before drawing a configuration conclusion.