Skip to main content
The /universes/{id}/brands and /universes/{id}/brands/{brand} endpoints expose a brand-centric view of your dashboard data. Instead of querying ai_visibility_ranking, then ai_competitive_analysis, then geo_competitor_analysis and joining them by brand_name yourself, the API does it for you.

​
What’s on a brand row

Every entry in the response carries seven numbers — three from AI mentions, two from GEO citations, plus rank and share-of-voice. Your own brand sits in the same array as every competitor.
​
string
Display name as it appears in the underlying history arrays. Preserves casing.Example: "verseodin.com", "Semrush", "tryprofound.com"
​
integer
Count of prompts where the AI mentioned this brand by name in its answer. Drawn from ai_visibility_ranking.Range: 0 to total_prompts for the universe.
​
number
This brand’s mentions as a percentage of total brand mentions in the universe. The share_of_voice values across all brands sum to ~100% (rounding aside).Example: 18.7 means this brand accounts for ~19% of brand mentions across the universe.
​
integer
1-indexed rank among all tracked brands by trust_mentions. 1 is the most-mentioned brand. Ties are broken by alphabetical order of brand_name.Tip: if your brand’s rank goes from 3 → 1 over a week, that’s the win you want to track.
​
number
Percentage of prompts where the AI mentioned this brand. Drawn from ai_competitive_analysis.coverage_pct.Difference vs trust_mentions: ai_coverage_pct is the rate (mentions / total_prompts × 100); trust_mentions is the count.
​
number
Percentage of prompts where the AI cited a URL belonging to this brand’s domain. Independent of whether the brand was named in the prose. Drawn from geo_competitor_analysis.coverage_pct.
​
integer
Count of prompts where this brand’s domain was cited at least once. Drawn from geo_competitor_analysis.prompt_count.

​
Three things this view answers in one call

​
”How is my own brand doing?"

​
"How is competitor X doing?”

Same call, just swap the brand name:

​
“Show me the leaderboard”

The list endpoint sorts by visibility_rank ASC — top of the array is the most-mentioned brand:

​
Filtering

Both endpoints accept the standard ?day= and ?engine= query params:
If multiple history rows match the filter (e.g. you don’t pass engine= and the universe has multiple engines), the most recent one is used. The response’s day + engine fields tell you which row was actually consulted, so you can detect this.

​
When a brand isn’t found

The single-brand endpoint returns 404 not_found when:
  • The brand name doesn’t appear in any of the three source arrays for the matched history row, OR
  • No history row matches the day/engine filters
The list endpoint returns 200 with data: [] and day: null, engine: null if the filters match no history row at all.

​
How this maps onto the raw history columns

If you want to do this synthesis yourself instead of using these endpoints, here are the three source columns: Match brand_name case-insensitively across the three arrays.