I'm afraid that a moderation API alone would not be enough of a measure against spammers. It presumes that automod applications will be developed, that admins will know about them, and that admins will set them up. Which would require self-hosting because automod-as-a-service would have awful privacy implications.
That's a shame though because experimentation over a moderation API would be more effective than a centrally designed upstream solution.