# ID Capture
ID Capture is a core onboarding step. It collects high-quality images of a user’s government-issued ID. For two-sided IDs, both the front and back are captured. The module also extracts data using optical character recognition (OCR) and runs authenticity checks.
***
## Where it fits in the flow
In an onboarding flow or workflow, **ID Capture** usually appears immediately after consent and document type selection and before Selfie Capture and final submission.
Once the user provides valid images, the flow continues to data extraction, document liveness/authenticity checks, and any downstream logic (such as Face Match with later selfie).
***
## User experience flow
The experience guides users from preparation to a successful ID capture. After granting camera permission, users see a short tutorial. During capture, the system evaluates framing, glare, blur, and edge alignment in real time. When requirements are met, the ID is auto-captured. If auto-capture doesn’t trigger in time, manual capture is available. In this case, images are then uploaded for quality checks with retry options on failure.
***
## Full Flow Map
This diagram shows all **ID Capture** screens, from tutorial and permissions through front/back auto/manual capture, uploading, and feedback (success/error).
***
## Happy Path (Light & Dark)
The ideal user journey occurs when the ID is captured successfully with no interruptions.
The happy path represents the smoothest version of the experience:
* The user grants camera access
* The user follows the tutorial
* The system detects proper alignment and lighting
* The ID is captured automatically without requiring retries or manual intervention.
In these images, both light and dark mode previews are included so teams can validate visual consistency across themes.
***
## Best practices for design and implementation
**✅ Do**
* Keep instructions short and actionable (for example, “Place your ID on a flat surface. Avoid glare”).
* Use a high-contrast, uncluttered background. Ensure text remains readable.
* Guide with real-time feedback, like, “Move closer,” “Avoid glare,” or “Show all corners.”
* Provide retry options and clear error messages (for example, “Wrong document side”).
* Respect accessibility with voiceover labels, adequate tap targets, and captions.
**❌ Don’t**
* Don’t reduce overlay opacity so much that the capturing frame edges become hard to see
* Don’t rely solely on color for status. Add icons/text.
* Don’t skip critical error states like glare, blur, or cut-off).
* Don’t let UI elements obstruct the camera area.