What's New

KyedNS now speaks the Model Context Protocol, the standard that lets AI assistants like Claude use external services. Point your assistant at kyedns.com/mcp with the API token from your account settings, and it can add, edit and deploy the DNS records on your domains, check monitoring, and look after your hostnames, all in plain language. A few things stay in the dashboard for now: DNSSEC, record templates, webhooks and monitors, and adding or deleting a domain. Every change made this way is marked in your activity log, and account settings now has a button to regenerate your API token, which instantly cuts off any client using the old one. Because tokens could previously only read your account and can now also make changes, every existing API token has been replaced as a precaution: if you use the API, grab your new token from account settings. Setup details are on the API page.

A webhook delivery that hits a server error or timeout is now retried a few times instead of being dropped, so a short restart of your endpoint no longer loses events. And a failing webhook is no longer disabled permanently after three failures: it is paused with a growing cooldown, you get an email about it, and the first successful delivery resets everything — no manual re-enabling needed.

Records that can be monitored now show an Uptime monitoring panel on their edit page — pick a port and start monitoring without leaving the record. The same panel appears on a free hostname's settings page, shows the current status once a monitor exists, and lets you pause or remove it. Monitored records also get a status pill in the domain's records list.

A host is now checked a second time before it counts as down. A single lost packet on the way to your host used to be enough to trigger a "Host down" email, and because the next check only runs after your chosen interval, the recovery mail then reported a full hour of downtime that never happened. Only a host that fails both checks is reported as down. Monitoring is also part of your activity log now: starting, pausing, resuming and stopping a check are recorded, as is every time a host goes down or comes back. The entries appear on the Activity page and on the zone or domain the host belongs to, so you can see the full history rather than just the last 24 checks.

TXT record values now accept a DKIM key copied straight from your mail server's key file or a dig answer — including the record name, parentheses, quoted line chunks, and trailing comment. We unwrap the formatting and store the key itself, so a blind copy-paste just works. Previously the quotes could survive into the record and break DKIM validation.

The Dynamic DNS panel on the dashboard checked the wrong field for dynamic records on custom domains, so a record with a perfectly good IP address could show as "No IP set" and count against the active total. It now looks at the address the record actually serves.

Older MX records stored the priority as part of the mail server value, so the edit form showed something like "25 mail.example.com" in the Mail Server field with the wrong priority beside it. These records are now stored the same way as new ones, and the edit form shows the priority and mail server in their own fields.

DKIM generators and DNS tools often present long TXT values as several quoted parts. Pasting that format as-is could previously make the record fail to resolve. The parts are now joined automatically, so both plain and quoted formats work.

The propagation checker now takes record names the same way the record forms do: enter _dmarc or mail._domainkey and it is checked under your domain automatically. Long record values in tables also hint that a click expands them.

Long values such as DKIM keys squeezed the record tables into unreadable columns, especially on mobile. They are now shown collapsed to a few lines — tap or click a value to expand it. Records with special characters are also now always published to DNS in correctly sized strings.