| Web.de | |
|---|---|
| Domains | web.de |
| Mail routing | mx-ha02.web.de and mx-ha03.web.de at the time of writing, September 2026. Match on the mx-ha pattern and not on one numbered host. The operator is shared with GMX, though under a separate brand. |
| What we can resolve | We resolve most mailboxes here |
How Web.de answers a probe
What the server does when we open a connection and ask about one address.
- Unknown local parts return a 550 with consistent wording, and the verdict is reliable.
- Connection limits are strict in the same style as GMX, since the same company sets them.
- Web.de sells a paid tier, and paid accounts are more likely to be actively used than the free ones.
- Umlauts in local parts show up regularly and need correct handling before the probe runs.
The part that catches people out
What we do about it
We treat United Internet's ranges as one budget across both brands, which keeps the answers flowing on German consumer lists where the two together can be a third of the addresses.
Sending to Web.de mailboxes
Verification is one step. These are the things that decide whether the message lands.
- German consent documentation matters, and this audience is more likely than most to act on a missing opt-in.
- United Internet's filtering responds quickly to complaints and forgives slowly.
- Web.de and GMX together dominate German free webmail, so a German list that shows neither is worth questioning.
Reading the results on a German shop list
Say an online bike shop in Cologne sends you 9,000 customer addresses, and 1,500 are on web.de. The run comes back with most of those resolved and 140 marked unknown, each with throttled as the reason. Here's how to read that. The resolved rows got a direct yes or no from the server, and the invalid ones can be suppressed today. Nothing about the 140 is suspicious. They were simply still in line when Web.de decided our connection had asked enough questions for now.
Because unknowns aren't billed, the sensible move is to export just those 140 and run them again tomorrow. Don't mail them in the meantime on the theory that they're probably fine, and don't delete them on the theory that they're probably not. Your coverage figure for the list should go up after the second pass, and that's the number to show your client.
De-Mail addresses are a different system
German lists sometimes contain addresses ending in web.de-mail.de, gmx.de-mail.de or t-online.de-mail.de. They look like a typo with an extra bit glued on. They're De-Mail addresses, from Germany's closed system for legally binding electronic letters, and ordinary internet email doesn't deliver into it. No SMTP probe from outside will get a meaningful reply, and no newsletter will ever arrive.
Don't repair them by trimming the suffix. name@web.de-mail.de and name@web.de are separate registrations, and the second may not exist or may belong to someone else. The clean fix is to pull De-Mail rows out before verification and ask those customers for a regular address the next time they order. While you're checking domains, note that web.de publishes DMARC p=quarantine (our lookup, September 19, 2026), so a customer-facing tool that sends "from" a web.de address on someone's behalf will land in spam.
Questions people ask
Is Web.de easy to verify?
Yes, relatively. It gives clear rejections for addresses that do not exist and does not run an accept-all anti-harvesting policy. Throughput is the only real constraint.
Is Web.de the same as GMX?
Same owner, shared infrastructure, separate brands and separate user bases. They behave nearly identically under a probe and share rate-limit budgets.
How should umlauts in an address be handled?
Encode them correctly before the probe and store them correctly in your list. A mangled local part fails for a reason that has nothing to do with the mailbox.