Brand sources
What are sources?
AI models (OpenAI, Gemini, Perplexity) often cite specific web pages in their responses. The Sources section in the brand dashboard shows which domains are associated with your brand in those responses — in other words, where AI “draws” information from when it mentions you.
The Sources view filters results only to mentions assigned to your brand. Competitor citations are not included — each domain is counted only if AI linked it to your brand name or one of its aliases.
KPI metrics
At the top of the Sources view, four aggregate indicators are shown for the selected date range:
| Metric | Meaning |
|---|---|
| Total citations | Total number of cited URLs in snapshots related to your brand |
| Unique domains | Number of unique domains (without duplicate URLs) |
| Source coverage | Percentage of snapshots with at least one citation of your brand (snapshots with a citation / all) |
| Top source type | Most common source type (for example, official_brand_content, news_media, user_forums) |
Domain table
The domain view presents a ranking of sites by citation count:
- Citation count — how many times the domain was cited in connection with your brand
- Unique snapshots — in how many different snapshots the domain appeared
- Avg. rank — average citation position in the AI response (lower = appeared earlier in the text)
- Source type breakdown — list of source types for a given domain with separate counters
Plan limits
| Plan | Number of visible domains |
|---|---|
| Free | 5 (the rest are hidden) |
| Starter | All |
| Enterprise | All |
On the Free plan, the table shows only the 5 most frequently cited domains. If the database contains more, the table displays the Plan limited marker.
Citation trend over time
A line chart shows the daily number of citations for the top N most cited domains plus the aggregate Others category (all remaining domains combined).
- Axes: date (X) × number of citations (Y)
- Domains selected dynamically (default: top 5 for the chosen period)
- The “Others” category appears only if there are domains outside the top N
URL preview for a domain (Starter+)
Clicking a domain row in the table opens a side panel with a list of specific URLs cited by AI for that domain.
For each URL, you can see:
| Field | Description |
|---|---|
| URL | Full page address (clickable) |
| Snapshot date | Date of the snapshot where the citation appeared |
| Rank in response | Citation position in the source list of the AI response |
| Source type | Source type |
| Proxy? | true if the URL is a vertexaisearch.cloud.google.com redirect (Google Grounding API) instead of a direct link |
Proxy URL means AI (Gemini) referenced the page through the grounding mechanism, but the system did not receive a direct link. The actual content corresponds to the domain shown in the table.
The panel loads up to 200 URLs (most recent chronologically).
Drill-down availability
| Plan | URL preview |
|---|---|
| Free | ❌ Locked |
| Trial | ❌ Locked |
| Starter | ✅ Available |
| Enterprise | ✅ Available |
How does brand filtering work?
Since the introduction of the entity_name field in the answer_sources table, each new snapshot assigns the cited page to a specific entity (for example, adidas, Nike, Under Armour). The Sources view for your brand shows only the URLs where:
entity_name IS NULL— records from before the update (legacy; displayed for backward compatibility)- or
entity_namematches the main brand name or one of its aliases (matched by lowercase + trim)
This ensures that pages cited about competitors do not pollute your brand view.
Choosing a date range
The date-range selector (from / to) at the top of the brand dashboard affects all Sources views. The default range is the last 7 days.