Document OCR

What it does

The OCR is the reading half of document verification. It detects the document in the photo, classifies it, and reads its fields into data your system can use. It reads printed and handwritten text, and it reads the MRZ with checksum. It also runs signature and stamp checks.

What is read

• Printed text. The fields printed on the document, such as the name, the date of birth, the document number and the expiry date.
• Handwritten text. Handwritten entries on the document are read as well.
• The MRZ. The machine-readable lines are read together with their checksum. The MRZ page explains what that means.
• Signatures and stamps. These are checked as part of the same pass.

Coverage

The OCR works across 10,000+ document types, 195+ countries and territories. The supported documents page lists the document families.

How it works in the flow

1. The user photographs the document in the widget.
2. The document is detected and classified.
3. Its fields are read, including the MRZ.
4. The anti-fraud checks run on the same images.
5. The result is added to the session.

What your system receives

A document entry in the session results, with the status of the step. See session statuses for every value.

Set it up

Document steps are part of the scenario. In the scenario editor you choose the documents and the countries you accept.

Next

Continue with document authenticity, or see the full flow in the interactive tour.