Skip to content

UX Recommendations

This section includes design and configuration recommendations for the Video SDK flow, which typically involves three phases: selfie capture, document presentation, and optionally, spoken identification. These suggestions are based on user testing and integration experience, aiming to reduce errors, improve clarity, and optimize completion time without compromising verification quality.

Video Capture Flow

  • The voice step (full name and surname) should be removed if not strictly necessary Reducing the number of actions makes the video shorter and more fluid. Unless the spoken identification is required for compliance or business validation, omitting this step can simplify the process and reduce abandonment.
  • It is recommended to keep the template overlay visible when capturing the back side of the document Since there is no intermediate screen indicating that the user must flip the document, retaining the template (GIF) inside the framing rectangle helps guide them visually. It reinforces the idea that a different side of the document must be shown and supports proper framing.
  • Maintain clear separation between actions within the video flow Even though all steps (selfie, document, voice) happen in a single continuous recording, visual instructions should clearly distinguish what is expected at each point (e.g., “Hold your document”, “Look at the camera”, “Say your full name”). This segmentation improves comprehension and pacing.

Tutorials and Instructions

  • Instructional screens should clearly explain what will happen during the video Users should know in advance that they will need to perform multiple steps: showing the front and back of a document, keeping their face visible, and optionally speaking aloud.
  • Images or icons should be action-oriented Use visuals that demonstrate what the user is expected to do (e.g., turning a document, showing both sides). Avoid decorative or abstract visuals that don’t directly reinforce the task.
  • Warn about camera and microphone permissions Letting users know that access to both the camera and microphone will be requested helps avoid confusion and increases the likelihood of consent.
  • Text should remain brief and aligned to the left Avoid long paragraphs or centered text. Use short phrases that are easy to scan and understand, especially under time-sensitive conditions.
  • Single instruction screen If there is only going to be a single instruction screen, it is advisable to indicate in the title that there will not be two steps. Additionally, it would also be advisable to change the button title, which should display text indicating to start the capture.
  • Keeping the close button visible is recommended Providing an option to exit maintains user control and avoids frustration in case of unexpected interruptions or confusion during the process.
  • Avoid altering detection logic or timing thresholds Modifying delays, timeouts or auto-capture behaviors can affect the precision and consistency of the verification process.
  • Error handling should guide correction In the event of a failed recording (e.g., document not visible, face out of frame), display brief, actionable tips to help users correct the issue without frustration.

Texts accesibility

  • Alt text allows people with visual impairments to understand the content of images using screen readers. Our SDKs support this functionality, and alt text can be modified in the settings.
  • However, two of the elements share one peculiarity: the spinner's alt text and the loading title tend to be very similar. That's why we recommend leaving the alt text blank if you decide to include a title, and vice versa, to prevent screen readers from reading two similar texts.
  • Please check this on the loading screen (loading.altText and loading.title) and on the camera loading screen in the capture (capture.cameraLoading.altText and capture.cameraLoading.title).