CN8 API Documentation

Detect and segment objects in images. Auto mode finds all objects; prompt mode targets specific objects via points or bounding box.

In auto mode, the service finds every object in the image, filters out tiny and low-confidence masks, and returns up to 100 segments sorted by area. In prompt mode, it predicts masks from the points or bounding box you provide and returns up to 3 candidates ranked by confidence. All masks are returned in COCO RLE format.

Segment Image

curl -X POST https://api.cn8.io/v1/proxy/studio-segment \
  -H "X-API-Key: <cn8_api_key>" \
  -H "Content-Type: application/json" \
  -d '{"imageUrl":"https://example.com/file"}'

{
  "status": "success",
  "data": {
    "image_size": { "width": 1920, "height": 1080 },
    "segments": [
      {
        "id": 0,
        "bbox": [120, 60, 380, 290],
        "area": 45230,
        "predicted_iou": 0.96,
        "stability_score": 0.93,
        "mask_rle": { "size": [1080, 1920], "counts": [0, 500, 200] },
        "centroid": [310, 205]
      }
    ],
    "segment_count": 12
  },
  "usage": { "units": 1, "unit_type": "item" }
}

Auto Mode

In auto mode, the model scans the entire image (or region if provided) and returns up to 100 segments sorted by area. Each segment includes a COCO RLE mask that can be passed directly to the Inpainting service.

curl -X POST https://api.cn8.io/v1/proxy/studio-segment \
  -H "X-API-Key: apk_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "imageUrl": "https://s3.../photo.jpg",
    "mode": "auto"
  }'

Prompt Mode

In prompt mode, provide point coordinates or a bounding box to target a specific object. Returns up to 3 mask candidates ranked by confidence.

curl -X POST https://api.cn8.io/v1/proxy/studio-segment \
  -H "X-API-Key: apk_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "imageUrl": "https://s3.../photo.jpg",
    "mode": "prompt",
    "points": [{"x": 450, "y": 300, "label": 1}]
  }'