VisionAI Object Detection

VisionAI uses a custom YOLO-v9-based architecture fine-tuned on 80 COCO classes plus 30 additional custom ones. Supports JPEG, PNG, WebP, and MP4. Includes confidence scores, bounding boxes, and segmentation masks. Available as both REST and WebSocket streaming.

Available Endpoints

const res = await fetch('https://api.visionai.dev/v2/detect', {
  method: 'POST',
  headers: { 'X-API-Key': process.env.VISION_KEY, 'Content-Type': 'application/json' },
  body: JSON.stringify({ image_url: 'https://example.com/photo.jpg', confidence: 0.7 }),
});
const { detections } = await res.json();
console.log(detections); // [{ label, confidence, box }]

API Access

API Provider

Nina Torres

Data & ML engineer turning models into production APIs.