Cumulative() with aggregate data?

I’ve got an experimental dataset with about 40 combinations of predictors and a 5-category Likert outcome. I’d like to model them with a cumulative logit likelihood. The number of cases is really large, so I’d ideally like to work with aggregate data giving the number of cases choosing each response option in each group (ala the | trials() interface for binomial and multinomial models). Is this currently possible in brms?

You could use the weighting function to multiply the likelihood of each predictor-outcone combination with the number of cases with this combination.

Ah, I didn’t realize that brms::weights() doesn’t normalize the weights like glm()! Thanks!