Run an IPv6 test from this browser
An IPv6 test checks whether a specific connection can use Internet Protocol version 6. This tool sends one request to an IPv4-only address endpoint and another to an IPv6-only endpoint. Run the test to see which requests succeed and the public address each endpoint observes. The IPv6 test runs in your browser, because a request made by our server would measure our server's connection instead of yours.
You do not need an account, a browser extension, or camera permission. The two requests run independently, and each has an eight-second timeout. The IPv6 test names the external endpoints before you run it: api.ipify.org for IPv4 and api6.ipify.org for IPv6. Those services receive the source address used to connect to them. Results describe the current browser path and should be refreshed after a relevant network change.
Before starting, note whether you are using Wi-Fi, cellular data, a work network, or a VPN. These conditions affect the route. If you are comparing configurations, run an IPv6 test before changing a setting and again afterward. Change one setting at a time. A before-and-after comparison is easier to interpret than several changes followed by a single result, especially when a device has more than one active interface.
Interpret an IPv6 test result
If the IPv6 request succeeds, the browser reached the IPv6-only endpoint during this measurement. The displayed address is the source observed at that endpoint. A successful IPv6 test is useful evidence of usable IPv6 on that path. It does not establish that every IPv6 destination is reachable, that every application uses IPv6, or that the connection has a particular speed or security level.
If both requests succeed, both address families were usable for these destinations. That is consistent with a dual-stack connection, although translation and other network mechanisms can complicate the complete architecture. The IPv6 test cannot inventory all of those mechanisms from two HTTP responses. Treat the measured reachability as the direct observation and any explanation of the wider network as something to confirm separately.
Why does the homepage show IPv4 when this IPv6 test works?
A website that supports both families can be reached over either. The browser and operating system choose a connection according to availability and timing. One request showing IPv4 does not mean the device lacks IPv6. Conversely, one IPv6 result does not establish that IPv4 is unavailable. This IPv6 test uses separate family-specific endpoints to make that distinction clearer than a single dual-stack request can.
What does “endpoint not reached” mean?
It means the relevant request did not produce a valid successful result within the test conditions. Possible causes include unavailable IPv6 routing, a DNS problem, a firewall, an extension, a VPN policy, or an endpoint outage. A failed IPv6 test should not be rewritten as “your device has no IPv6 address.” A device can have IPv6 configured locally while lacking a working route to this destination.
Do not treat a missing response as a privacy certification either. If a filter blocks the test domain, the IPv6 test may fail even though another IPv6 destination remains reachable. Repeating the same check once can rule out a transient failure, but a persistent problem needs independent evidence. Device interface settings, router status, and a provider's support information can help distinguish local configuration from an upstream issue.
What this IPv6 test does not measure
This IPv6 test measures browser HTTP reachability to named endpoints. It is not an ICMP ping, a traceroute, a speed test, or an exhaustive firewall test. The eight-second limit is a timeout, not a latency score. Browser scheduling, DNS resolution, connection setup, endpoint load, and network conditions all affect the request, so a single successful response should not be converted into a precise network quality rating.
The test also cannot inspect another application's route. A VPN may carry some applications while excluding others, or it may apply different policies to IPv4 and IPv6. Run the IPv6 test in the browser you actually want to assess. If a game, command-line tool, or managed work application behaves differently, collect evidence from that application's own connection instead of assuming the browser represents everything on the device.
Temporary addresses and changing results
IPv6 devices can use temporary addresses for outbound connections. Multiple addresses can also exist on one interface for different scopes or purposes. A changed IPv6 test result is not automatically a fault. Record the time, network, and VPN state before deciding whether the change is unexpected. A link-local address beginning with fe80 is a separate concept and is not the public address this endpoint observes.
NAT64 and translated connections
Some networks let IPv6 clients communicate with IPv4 services through translation. As a result, a working IPv4 destination does not always prove that the local interface has an ordinary public IPv4 address. This IPv6 test reports the family-specific endpoint observations without claiming to reconstruct the translation architecture. Your provider or network administrator can confirm whether DNS64, NAT64, or other transition mechanisms are used.
IPv6 test questions
Should I disable IPv6 if the test fails?
An unsuccessful IPv6 test is a reason to investigate, not an automatic instruction to disable a protocol. Disabling IPv6 can affect other services and may hide the original problem without resolving it. Check whether your router and provider support IPv6, whether the active interface has an appropriate address, and whether a VPN changes the result. Follow device and provider documentation when changing settings.
Can an IPv6 test reveal a VPN leak?
It can provide one piece of evidence. Compare the source seen with the VPN disconnected and connected, and understand the VPN's intended routing policy. An IPv6 test that succeeds while the VPN is active does not by itself prove a leak: the IPv6 address may belong to the VPN exit. Likewise, a failed request cannot prove all IPv6 traffic is blocked. A complete assessment needs known baselines and tests for other relevant traffic.
Why does an IPv6 address look so long?
IPv6 uses 128-bit addresses represented in hexadecimal groups. Consecutive zero groups can be compressed using ::, and leading zeros inside a group can be omitted. The IPv6 test displays a normalized form without shortening it to an ellipsis, so you can copy the exact address. Two differently written strings can refer to the same numerical address; normalize them before comparing values from different tools.
Can I use the result for a firewall rule?
You can use the observed address as evidence, but consider whether it is temporary and whether the rule should allow a single address or an assigned prefix. An IPv6 test does not authorize a firewall change or establish ownership of a whole range. Verify the allocation and use the narrowest rule that meets the intended need. The subnet calculator helps check a prefix's boundaries before you configure it.
How do I report an IPv6 test problem?
Include the failing family, observation time, active network, browser, VPN state, and whether another network produces a different result. Avoid sharing unnecessary account information. If the IPv6 test succeeds on cellular data but fails on home Wi-Fi, that difference narrows the investigation. If both families fail only for the test endpoint, check filtering or endpoint availability before assuming your entire internet connection is down.
Sources
The family-specific endpoint behavior is documented by ipify. IPv6 addressing is described in RFC 4291, normalized text representation in RFC 5952, and connection selection behavior in RFC 8305. These references explain mechanisms; the IPv6 test above supplies a current observation with a deliberately limited scope.
For an IPv6 test that does not return an address, follow the IPv6 not detected checklist to separate browser, network, router and VPN causes.
MyIP.dog