The API can return the same figures as the Attributes and Objections cards on the Perception page. For keys and how to send a request, see API reference.
Addresses
Both are GET, read only, and take from, to (as 2026-10-01) and engines (a comma separated list) like the other report addresses.
| Address | Same figures as |
|---|---|
/api/v1/projects/{id}/perception/attributes | The Attributes card |
/api/v1/projects/{id}/perception/objections | The Objections card |
An address answers only for a workspace that has the matching card. It follows the same keys, scopes, request limits and project limits as every other address, and a key needs no write permission.
What comes back
- Attributes:
rows(each topic with its claims, answers, share, positive, neutral and negative counts, tone, sentiment score, whether it is thin, and its rank among the tracked brands),knownFor(the card's Known for chips),split(the positive, neutral and negative split) andanswers(the card's Based on N answers). - Objections:
rows(each topic with its negative claims, the answers raising it,prevalence, the engines, up to two quotes and up to three cited sites),claimsandanswers. - Both also return
project,latestCheck,period,enoughand aunitsline.
Units
These two addresses use the app's own units so you can match them with the cards. share, prevalence and split run from 0 to 1, where the card shows percentages. score runs from 0 to 100. This differs from the other addresses, where rates are already percentages.
A date range with no completed check returns empty: true with the same message as the other reports.
Not in the API
The history of attributes, a written summary of strengths and weaknesses, and filters by market or tag are not included. Neither address has an MCP tool.
For the cards themselves, see Perception.