Training or filtering? A proper answer to this question requires its own blog post :)

Long story short, Mailgun does have a spam filtering system: if you enable it for your domain, we'll put all your traffic through a cloud of SpamAsassin machines. Those machines themselves need to be re-trained every once in a while and we do that ourselves without providing any APIs to control that process because, frankly, we believe we can do it better. The spam score along with "Mailgun spam verdict" is POSTed into your route handler - I will update the blog post to include this: thank you for reminding!

Thanks for the description.

Though one company's spam is another company's ham, so I wonder if this would work well enough across all your clients?

You're right, this is where per-mailbox spam filtering kicks in. But for that to work you need to create a programmatic mailbox (we have APIs for that) which allows us to collect your spam&ham datasets.

We do have a few customers who're using per-mailbox spam training via our "Beta API access" program, but we have not made the decision to launch it as a product yet. If this is something you may be interested in, reach me via email in my profile.

Thanks.