Vision›OCR Extract
API Key not detected. Click "Configure API Key" in the top-right corner before using the playground.
u1-ocr
*Input Image

Supported formats: JPG / JPEG, PNG, BMP, PDF

Size limit: Images ≤ 10MB, PDF ≤ 20MB

General Extraction
Task Info
Trace ID-
Status-
Extraction Result
Result will show here
Guide
This is a synchronous API: upload an image or PDF (u1-ocr only) with a prompt to get the extraction result immediately.
Write a precise prompt and specify the output format (e.g., JSON) for structured parsing. For PDF, content is a JSON string of a per-page array — JSON.parse it before use.
API Reference
CURL
curl -X POST "https://maas-api.unisound.com/v1/ocr/image/extract" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "model=u1-ocr" \
  -F "type=image" \
  -F "file=data:image/png;base64,..." \
  -F "prompt=Your task is to extract information from the image above and structure it into JSON.
Strictly follow this schema for extraction (if a field does not exist, return an empty string): {\"invoiceNumber\":\"\",\"issueDate\":\"\",\"items\":[{\"itemName\":\"\",\"unitPrice\":\"\",\"quantity\":\"\",\"totalAmount\":\"\"}]}
Please output JSON only. All values must be strings. Do not provide any explanation."