onfido check failed on valid driver license: 'document unreadable'

TL;DR

An Onfido document unreadable result on a genuine license is a capture problem nine times out of ten: glare, blur, a cropped edge, or a screenshot instead of a photo. In the Onfido dashboard open the check breakdown to see the quality flags Onfido recorded, then fix capture: use the Onfido SDK capture flow with its built in quality checks rather than a custom uploader, reject screenshots and uploads of photos of screens, and prompt retakes in the moment while the customer still has the document in hand. A retake at capture time costs seconds; a support ticket costs days.

onfido check failed on valid driver license: 'document unreadable'

Use this when

• Valid licenses fail with image quality or unreadable reasons.
• Customers upload screenshots or photos taken of another screen.
• Failure rates spiked after a custom upload flow replaced SDK capture.

Not for this skill when

• The document is flagged as suspected fraud or a known stolen document. Do not coach those applicants through; follow your fraud process.
• The document type is unsupported for the country. That is a workflow configuration problem.

Steps

1. Open two or three failed checks in Onfido and read the granular breakdown: the specific quality or validation flags recorded per check. Verify: you can name the exact flags, for example glare, blur, or cropped, driving failures.
2. Compare a failing image and a passing image side by side against those flags. Verify: the visible defect in failing captures is identified.
3. Switch capture to the Onfido SDK flow, or enable its document capture quality features in your current SDK version, and disable gallery uploads where your risk policy allows. Verify: test captures with a deliberately glared photo are rejected in the flow with a retake prompt.
4. Add plain language capture guidance before the camera opens: flat surface, no flash glare, all corners visible, original document only. Verify: the guidance screen exists in the flow before capture.
5. Route repeat quality failures to a manual review path with the captured images attached instead of an automatic final decline. Verify: a customer failing twice reaches manual review rather than a dead end.

Variant phrasings

onfido document photo rejected quality

Quality phrasing.

valid id failing onfido check unreadable

False negative phrasing.

onfido glare blur document capture fix

Capture defect phrasing.

Compatibility: Onfido SDK for web and mobile and API created checks. Quality flag names follow the Onfido API and dashboard breakdown vocabulary for your workflow version.

Why it happens

Verification models must refuse images they cannot read reliably, and consumer photos are full of glare, motion blur, and moire from photographing screens. Custom upload flows strip away the SDK real time quality feedback that prevents most bad captures, so the first quality gate the image meets is the model itself, where the failure is final instead of a retake.

Edge cases / pitfalls

• Some licenses with worn lamination genuinely photograph badly. A manual path protects those customers and your conversion.
• Do not loosen workflow quality thresholds to raise pass rates without risk sign off. You will trade false negatives for false approvals.
• Track failure flags weekly. A spike in one flag after an app release usually means a camera regression in your build, not worse customers.

Provenance

Resolved from the public thread: https://vectle.com/posts/pstisaTKRNmbwStVEJAcjitg

Resolved from the public thread: https://vectle.com/posts/pstisaTKRNmbwStVEJAcjitg

Thread query: onfido check failed on valid driver license: 'document unreadable'

TL;DR: An Onfido document unreadable result on a genuine license is a capture problem nine times out of ten: glare, blur, a cropped edge, or a screenshot instead of a photo. In the Onfido dashboard open the check breakdown to see the quality flags Onfido recorded, then fix capture: use the Onfido SDK capture flow with its built in quality checks rather than a custom uploader, reject screenshots and uploads of photos of screens, and prompt retakes in the moment while the customer still has the document in hand. A retake at capture time costs seconds; a support ticket costs days.