biofrq.detect

Find faces in a single image. Returns bounding boxes and detector confidence. The cheapest building block — used on its own for crowd counts, and as the front half of every other action that needs to localize a face first.

Params

Field	Type	Required	Default	Notes
`images[0]`	`ImageItem`	yes	—	One image. See Envelope → Image inputs.
`maxFaces`	int	no	unlimited	Cap the returned face count. Must be > 0.
`returnAlignedCrop`	bool	no	`false`	Include a normalized 112×112 crop per face (base64 JPEG). Adds bytes; off by default.

Request

json

{
  "action": "biofrq.detect",
  "params": {
    "images": [
      {
        "kind": "image",
        "bytes": "<base64>",
        "contentType": "image/jpeg"
      }
    ],
    "maxFaces": 5,
    "returnAlignedCrop": false
  }
}

Response

json

{
  "ok": true,
  "requestId": "...",
  "action": "biofrq.detect",
  "data": {
    "results": [
      {
        "index": 0,
        "correlationId": "auto-0",
        "ok": true,
        "value": {
          "faces": [
            {
              "faceBox": { "x": 142, "y": 87, "w": 218, "h": 218 },
              "confidence": 0.992
            }
          ]
        }
      }
    ]
  },
  "errors": []
}

An empty faces array is still a successful response — "no face found" is a valid detection outcome and is billed.

Common errors

INVALID_REQUEST — missing or malformed params.images[0], maxFaces ≤ 0.
PAYLOAD_TOO_LARGE — decoded image exceeds the size cap.

Open in playgroundbiofrq.detect