Neural Shelf Intelligence
Detect, classify, and count every SKU on every shelf with enterprise-grade computer vision. Sub-second inference on proprietary YOLO11 models trained on millions of real-world retail images.
POST /api/v1/inference/
{
"status": "completed",
"processing_time_ms": 145,
"result": {
"detections_count": 47,
"confidence_avg": 0.9432
}
}Engineered for Retail Scale
A complete computer vision stack designed for shelf audits, planogram compliance, and share-of-shelf analytics.
Proprietary YOLO11 Models
Custom-trained on millions of real-world shelf images for unparalleled accuracy in product detection and classification.
TensorRT Inference
GPU-accelerated workers deliver sub-200ms predictions with dynamic batching and result caching.
SKU & Category Mapping
Automatic mapping from class IDs to SKUs and hierarchical categories with quantity counting.
Developer-First API
RESTful API with granular API keys, webhooks, rate-limiting, and usage analytics built for scale.
Real-time Analytics
Track inference volume, confidence scores, credit usage, and processing times across your organization.
Enterprise Security
Multi-tenant organizations, role-based access, audit logging, and encrypted storage for every image.
Async Inference Pipeline
Queue-based, non-blocking, and fault-tolerant. Built on Celery + Redis with automatic retries and circuit breakers.
Upload Image
Send a shelf photo via URL or multipart upload. We validate MIME types, magic bytes, and plan limits instantly.
GPU Preprocess
Images are decoded, resized with letterboxing, and normalized on GPU workers to minimize CPU overhead.
TensorRT Inference
ONNX engines execute asynchronously on TensorRT with NMS and dynamic batching for maximum throughput.
JSON Results
Receive structured detections with bounding boxes, SKU codes, confidence scores, and hierarchical categories.
Built for Developers
Drop an image, get structured JSON. Integrate in minutes with our REST API, webhooks, and granular API key scopes.
# Create a job
curl -X POST https://api.retail-ai.shop/api/v1/inference/ \
-H "X-API-Key: sa_live_..." \
-F "image=@shelf.jpg" \
-F "callback_url=https://myapp.com/webhook"
# Response
{
"id": "550e8400-e29b-41d4-a716-446655440000",
"status": "queued",
"credits_used": 1,
"estimated_wait_seconds": 2
}Simple, Credit-Based Pricing
No hidden subscriptions. Pay for what you process. One credit = one image analyzed.
Free
Perfect for prototyping and small pilots.
- 512px at 1 credit
- 1024px at 5 credits
- 2048px not available
- Standard model access
- Community support
Starter
For growing teams with real audit workloads.
- 512px at 1 credit
- 1024px at 5 credits
- 2048px not available
- 60 RPM / 5,000 RPD
- Webhook support
Pro
High-throughput CPG and retail analytics.
- 512px at 1 credit
- 1024px at 2 credits
- 2048px at 5 credits
- 300 RPM / 50,000 RPD
- Priority GPU queue
Enterprise
Custom SLAs, multi-region, and white-glove onboarding.
- 512px at 1 credit
- 1024px at 2 credits
- 2048px at 2 credits
- Multi-GPU allocation
- SSO & SCIM
- 24/7 phone support