Local AI · Vision & OCR
Retail Label Extractor
Live photo capture · Image to structured JSON
Checking service...
01 / Connection
Model settings
OpenVINO · Intel CPUSelect an installed vision model. Only administrators can view or modify these settings.
Recommended for displaying every detected key/value pair.
512 is a faster starting point on 8 Intel CPU cores. Use 768–1024 for tags with many details.
Checking connection. On first launch, monitor the model download and load with
docker compose logs -f openvino.02 / Instructions
Image analysis prompt
Editable · Full JSONThe API returns flexible field names. The result viewer automatically expands nested JSON and arrays into separate key/value rows.
01 / Input
Image source
Upload or live webcamDrop an image or click to uploadJPEG, PNG, WebP · Maximum 12 MB
No image selected
Live webcam capture
Photo · No audio
Camera live
Allow camera permission when asked. Browser live preview requires HTTPS or localhost; on a phone over HTTP, use Phone camera instead.
Ready. Upload an image, capture a photo, or use the sample tag.
02 / Output
Extraction results
0 fieldsReady for analysis
Extracted key/value pairs from your image will appear here alongside the original image.
| Key | Value |
|---|