Run guardrail versions. Accepts application/json (text-only) or multipart/form-data (with optional images, up to MAX_IMAGES).
| Time | Status | User Agent | |
|---|---|---|---|
Retrieving recent requests… | |||
Runs one or more guardrail versions on your input and returns a result for each rule. Send application/json for text only, or multipart/form-data to include images. A request that is too large returns 413.
When to use it
Use it to check content with a published guardrail directly from any AI application, inside or outside a SUPERWISE agent.
Required parameters
This call takes no required parameters.
Response
A successful call returns 200 OK with a JSON array of rule results. Each has the rule name, its type, whether the input is valid, and a message.
Related: List guardrail versions, Create a guardrail version, Run guardrules
Unauthorized
Not Found
Request Entity Too Large
Unprocessable Entity
Internal Server Error