Update methods

Four ways to tell us the address changed. All end at the same place: checked against the hostname's token, recorded in your history, published to DNS.

Method	Best for	Guide
Our script on the device	A Pi, a home server, anything that runs cron	Client guides
Router's built-in DynDNS	Most homes - set it once in the router	Router guides
The update API	Your own scripts and programs	Update API
The panel, by hand	Testing, one-off fixes	Panel guide

One hostname, one updater

Run exactly one updater per hostname. Two devices publishing different addresses for one hostname will fight, and the hostname will flap. If the updating device changes, turn the old updater off first.

IPv4 and IPv6 are independent

Each update can carry an IPv4 address, an IPv6 address, or both; each half is handled on its own. Sending only IPv4 never clears your IPv6 record, and the other way around. Or send no address at all: we publish the address the request came from (see the API reference).

How often should you update?

Only when the address changes. Every method here checks first and sends only on a change - routers do this themselves, and our scripts do it too.

Rate limits (see the API reference): 300 attempts per hour per source address, 120 authenticated updates per hour per account. Normal use never comes close; a client retrying in a tight loop can hit them, and the fix is to slow the client down.

What you see back

Every attempt is on your history page: time, source, what it tried to set, whether it worked. If an update seems to have no effect, look there first - the answer codes are explained in Troubleshooting. If your provider rotates an IPv6 prefix, IPv6 prefix mode covers the setup.