Skip to main content

Managing Long Receipts

The BlinkReceipt Scan API supports multi-frame processing for long receipts, enabling you to submit multiple images of a single receipt. Once the last frame is submittied, results will be aggregated and returned in the last request. We recommend allowing users to take multiple images of a long receipt in sections, rather than attempt to capture a long receipt from within one image. Proximity to a receipt will improve OCR results

Follow these steps to manage long receipts:

Step-by-Step Instructions​

1. Submit the First Frame:

  • Include the first image of the receipt.

  • Set frame_idx to 1

  • Set last_frame to false (indicating that more frames will follow).

  1. Note: Save the blink_receipt_id from the response, as it is required for subsequent frames.

2. Submit Subsequent Frames:

  • Include the blink_receipt_id returned from the first request.

  • Increment frame_idx by 1 for each new frame.

  • Keep last_frame set to false for intermediate frames.

Sample Request for Frame 2:

  1. Submit the Final Frame:

  • Include the blink_receipt_id from the initial request.

  • Set last_frame to true to indicate the end of the receipt submission.

  1. Sample Request for the Last Frame:

Key Notes:​

  • Order Matters: Ensure frames are submitted sequentially, starting with frame_idx=1.
  • Use Last Frame Flag: The final frame must have last_frame=true to indicate the completion of the submission.
  • Optional Final Call: If your workflow doesn’t know the last frame during real-time processing, make an additional API call after all frames are submitted with just the license_key, blink_receipt_id, and last_frame=true to finalize the session.