avclabs MCP

@avclabs.ai/media-mcp

Node.js >=18 · MIT License

Image Segmentation (SAM3)

<>sam3_predictSAM3 image segmentation (supports local path, URL, or Base64 image)

Use SAM3 segmentation API to analyze images and generate inference results (masks, boxes, scores).

Input options

ParameterTypeRequiredDescription
imagePathstringNoAbsolute path to local image. Supports common formats (PNG, JPG, JPEG)
imageUrlstringNoPublicly accessible image URL. Must be publicly accessible, does not support links requiring login or signature
imageBase64stringNoBase64 encoded image data. Large images will have large base64 data, transfer may take longer
promptstringYesEnglish text prompt specifying the target object to segment in the image. SAM3 model only accepts English prompts

Returns:

FieldTypeDescription
masksarray2D array. Each element is a binary mask (values 0 or 1) with the same dimensions as the input image, marking the pixel-level position of detected objects in the image. The i-th mask in the array corresponds to the i-th detected object instance.
boxesarray2D array. Each element is a bounding box coordinate in [x1, y1, x2, y2] format, representing the rectangular region of detected objects in the image. Coordinate system: origin (0, 0) at top-left of image, x-axis grows right, y-axis grows down, units in pixels.
scoresarray1D array. Each element is the confidence score for the corresponding detection result, ranging from 0 to 1. Higher scores indicate greater model confidence.

Request Example

json
{
  "imagePath": "/home/user/images/cat.jpg",
  "prompt": "find all cats"
}

Response Example

json
{
  "masks": [
    [[0, 0, 1, ...], [0, 1, 1, ...], ...],
    [[0, 0, 0, ...], [0, 0, 1, ...], ...]
  ],
  "boxes": [
    [120, 80, 300, 450],
    [400, 200, 600, 500]
  ],
  "scores": [0.95, 0.87]
}

Truncated return example (not completed within 50s)

json
{
  "success": true,
  "status": "processing",
  "task_id": "xxx",
  "message": "Task is still processing (waited about 50 seconds). Please retry later or record this task_id for manual follow-up.",
  "note": "The synchronous wait for this long-running task has been truncated."
}

FAQ → FAQ

avclabs

AI video enhancement service platform based on the MCP protocol, providing developers with powerful video processing capabilities.

© 2026 AVCLabs Inc. All Rights Reserved.