/simulate-review endpoint takes a user persona and a set of restaurant features, then returns a contextually authentic Nigerian restaurant review. The persona drives the reviewer’s tone, price sensitivity, and rating tendency, while the features shape what the review focuses on. Most integrations supply a persona object directly; the user_id field is an alternative that resolves a persona from the built-in review dataset.
Request
user_id or persona — not both, not neither. The API returns 422 if neither field is present.
Body
An ID from the built-in review dataset. The API resolves the full persona automatically. Use this or
persona, not both. Most integrations should use persona directly.A persona object supplied directly. Use this or
user_id, not both.The name of the restaurant or food item being reviewed (e.g.
"Yellow Chilli Victoria Island").The type of item. Typically
"restaurant".The city or area where the restaurant is located (e.g.
"Lagos").A list of attributes describing the restaurant’s experience (e.g.
["expensive", "nice ambience", "average food", "fast service"]). These shape what the generated review focuses on.Response
An integer star rating from 1 to 5, consistent with the persona’s rating tendency and the provided features.
The generated review text, written in a tone matching the persona. May include Nigerian Pidgin English phrases.
Describes the language style of the generated review. One of
"pidgin-heavy", "mixed", "formal", or "casual".Example
Errors
| Status | Cause |
|---|---|
422 | Neither user_id nor persona was provided, or required persona fields are missing. |
502 | The Gemini AI layer returned an unexpected or unparseable response. |
503 | The Gemini AI service is unavailable or not configured. |