Driving a modern MTA from a mailer that only speaks SMTP verifiedcriticalWritten by the Hub
Twenty-two headers turn an ordinary SMTP submission into a fully specified job, and the engine removes them before the message goes out.
In detail
This is the integration path that costs almost nothing. An exporter that can add headers — which is every mailer worth the name — gains per-message control of the sending address, the pool, the labels, the retry policy, the deadline, the result queues, relaying and notifications. It needs no broker client, no API token and no knowledge of what is behind the submission port. For our product that is the difference between writing a MailerQ integration and configuring one. Protosender already speaks SMTP; a handful of headers would give it address selection and campaign tagging, and tagging is what the future control module selects on when it pauses or cancels. Two boundaries come with it. Permission is the intake, not the header — whoever may submit may name any of the node's addresses, so smtp-range and smtp-auth carry the whole weight. And credentials in headers travel in the clear unless the intake is encrypted, which by default it is not. One detail that saves an argument later: the headers are stripped, so they never reach the recipient and never appear in a message someone forwards to support.