Account

What new tool do you want?

Developer Tools

DNS Lookup (A, MX, TXT, CNAME, NS…)

Look up DNS records for any domain — A, AAAA, CNAME, MX, TXT (SPF/DKIM/DMARC), NS, SOA, CAA, SRV and reverse PTR — compared across Google and Cloudflare to spot propagation issues.

How to use

  1. 1Enter a domain (or an IP for reverse lookup).
  2. 2Pick a record type, or press All records.
  3. 3Compare the answers from Google and Cloudflare DNS.

Frequently asked questions

Why do the two resolvers disagree?

DNS changes spread as cached records expire (the TTL). Right after a change, one resolver may still serve the old value — wait for the TTL to pass.

How do I check SPF, DKIM and DMARC?

Choose TXT. SPF is a TXT record on the domain; DMARC is on _dmarc.yourdomain.com; DKIM is on selector._domainkey.yourdomain.com.

What does DNSSEC validated mean?

The answer was cryptographically verified by the resolver, so it can't have been tampered with in transit.

A public resolver's view, not your ISP's

Lookups are sent as DNS-over-HTTPS requests straight from your browser to Google's (dns.google) and Cloudflare's (cloudflare-dns.com) public JSON APIs, run side by side so you can compare them. That's a deliberately different vantage point from whatever resolver your own ISP or office network is using — which matters when someone says a DNS change "isn't showing up for them": their local resolver may simply be serving a cached answer from before the change, independent of what either public resolver now returns.

Worked example: example.com's A and TXT records

Looking up A records for example.com currently returns 104.20.23.154 and 172.66.147.243 (the domain now sits behind Cloudflare's network), with both resolvers' results sorted and compared — matching results show a green "Both resolvers agree" badge. Switching to TXT surfaces "v=spf1 -all", an SPF record that explicitly authorizes no mail servers to send as that domain, plus a separate site-verification TXT string — exactly the kind of record the EXPLAIN panel describes TXT as holding (SPF, DKIM, DMARC, verification codes).

A reverse lookup secretly queries a different name

Choosing PTR and entering an IP doesn't query the IP directly — reverseName() first rewrites it into the in-addr.arpa format DNS actually uses for reverse lookups, so 104.20.23.154 becomes a query for 154.23.20.104.in-addr.arpa behind the scenes (IPv6 addresses get a similarly rewritten, nibble-reversed ip6.arpa name). You never have to construct that string yourself — type the IP and the tool handles the conversion.

When the two resolvers disagree

A mismatch between Google's and Cloudflare's answers almost always means a change is still propagating: each cached record expires according to its own TTL, and a resolver that queried the domain right before an update keeps serving the old value until that TTL runs out. The DNSSEC validated badge is a separate signal — it means the resolver cryptographically verified the answer came from the real authoritative server and wasn't tampered with in transit, which is unrelated to how fresh the data is.

CAA records: a lock few domains ever check

Of the ten record types supported — A, AAAA, CNAME, MX, TXT, NS, SOA, CAA, SRV and PTR — CAA is the one most site owners have never deliberately set, and checking it is worth doing specifically because an absent CAA record means any public certificate authority is allowed to issue an SSL certificate for that domain, while a present one restricts issuance to only the authorities you name. A domain with no CAA record isn't broken — it's simply relying on the default, open policy — but for anything security-sensitive, adding one closes off a real (if narrow) path to a mis-issued certificate.

What the TTL column next to each answer is for

Every returned record carries its own TTL in seconds, shown in the results table, and it's the mechanism behind the 'why do the two resolvers disagree' question: a record with TTL 3600 can be cached by a resolver for up to an hour after it last fetched it, so changing a record and checking immediately can show the old value simply because the previous TTL window hasn't expired yet anywhere it was cached. Lowering a record's TTL in advance of a planned change — many admins drop it to 300 seconds a day ahead of a migration — is the standard way to make a future cutover propagate faster.

Related tools

More: Developer Tools