Structured extraction
Find key product-label fields and bring the details together.
READ TEXT FROM PRODUCT IMAGES
Read text from product photos and organize origin, ingredients, manufacturer, shelf life, and specifications. Each field links back to its place in the image, so you can review anything uncertain.
SINGLE
ORIGIN
START WITH A PRODUCT IMAGE
Or drop images here, or⌘/Ctrl Vpaste · Multiple images supported
JPG / JPEG / PNG / WEBP · Up to 12 MB each · 16 megapixels max
Start with the details that matter
01 / WHAT YOU GET
Built for product detail images. See what it says—and where it came from.
Find key product-label fields and bring the details together.
Jump from any result to its place in the original image.
Keep the full OCR text, coordinates, and structured JSON for the next step.
02 / A CLEAR WORKFLOW
Review one image at a time. No guesswork, and no hidden evidence.
Upload
Drop, choose, or paste
Read
Keep text locations
Extract
Origin, ingredients, specs
Verify
Jump to the image
Export
JSON with traceable evidence
03 / EVIDENCE FIRST
Values, status, source image, original OCR text, and coordinates stay together. Missing means missing; uncertain fields are marked for review.
/api/ocrmultipart · file{
"origin": {
"value": "日本",
"status": "found",
"image_id": "image_2",
"evidence": [{
"text": "原产国名:日本",
"confidence": 0.98,
"box": [[62,410],[318,410],[318,446],[62,446]]
}]
},
"ingredients": { "status": "review", ... },
"images": [...],
"ocr_requests": [...]
}MAKE THE LABEL LEGIBLE