Each analysis carries its current run, scores (including brand share of voice), signed change versus the previous run, prompt count and recommendation count.

range defaults to daily: the newest completed run on or before yesterday in the analysis timezone, against the day before. Analyses without a completed run for the range are omitted.

Every filter is optional and narrows the result. The list is not paginated; pass ids to bound it.

providers recomputes scores, share of voice, and their changes from those AI providers only. Brand mention rate is the average of the selected providers; citation rate, sentiment, and share of voice pool their responses. Rows stay. A side that cannot be scored for the selection has no score and no share of voice. Brand Accuracy stays the pooled value.