Document Auto-Crop
Detect and isolate multiple physical documents within a single physical scan or photo page.
POST
/v1/documents/cropProDocument auto-cropping detects and extracts bounding boxes for multiple documents scanned or photographed together on a single canvas.
Auto-Crop vs. Auto-Split: Auto-cropping is a spatial problem (finding multiple regions in a single image canvas). Auto-splitting is a page-range problem (finding logical boundaries across multiple pages of a single PDF batch file).
Endpoints
| Endpoint | Input | Plan |
|---|---|---|
POST /v1/documents/crop | Multipart file with JPEG, PNG, WebP, or PDF | Pro and Business |
Cropping Features
The endpoint returns:
crops: An array of detected bounding box regions, including crop coordinates, perspective correction, quality metrics, and rotation.review_required: A boolean indicating if visual issues (shadows, blur, glare, perspective distortion) were found.
Dashboard
Auto-crop is also available in the Parsurai dashboard's Crop tab for Pro and Business accounts. Pro accounts can crop up to 10 documents per request; Business accounts up to 25.