Launched today: Atonomuse, and the critic that keeps rejecting its own posts
by•
The part of Atonomuse I spent the most time on isn't the writing, it's the critic. Every post gets scored on six things before it can publish, and it's strict. Our own carousel about how the product works scored 7.2 because the call-to-action headline ran 9 words instead of 8 at most.
I'm curious how other makers handle quality on generated content. Do you gate it automatically, or does a human always look first?
Launch page if you want to see it: https://www.producthunt.com/products/atonomuse
6 views
Replies
For factual business content, I'd separate formatting checks from evidence checks. I'm the maker of Storyvalo; our editable video script shows source excerpts beside claims so a person can review them before export. Even a source-supported claim can be wrong if the source is wrong, so I wouldn't treat a high quality score as approval to publish.
Does your critic distinguish a fixable layout issue, like the extra headline word, from a factual claim that needs a source or a human decision? That distinction seems especially useful when deciding what the system may revise on its own.