After playing around with it for a bit, it’s clear that #Stalwart isn’t ready for production. I think it will get there eventually, and it’s no slight against the team: writing an email server is hard. But for me, I’ll stick with the old standbys for now (Dovecot, Postfix). Stalwart is promising, with a modern design in a modern language, powerful scripting support, etc.

I thought I’d set it up for one of the domains I don’t currently use for anything, but even then I cannot seem to get it working. The support team is responsive, but if it’s this hard to get a user added so that they can log in, I don’t even want to know what other problems I’ll have.

I’ve spotted a few other gotchas, like using SHA-512 for password hashing (other, safer algorithms are supported, but this is the format used by default). The choice of RocksDB as a default database isn’t a particularly great one either, although you should be able to use other databases but I couldn’t get it to work with Postgres.

Again, I think that the team will get past these challenges and I look forward to a modern mail server. I’m just not ready to submit my users to it yet.