Assistive software that translates on-screen content into synthesized speech or braille output for users who are blind or visually impaired. Designing for screen readers requires semantic HTML, meaningful alternative text, logical reading order, and fully keyboard-navigable interactions.
Common contexts
- Testing a newly built form flow with VoiceOver on iOS to verify that field labels, error messages, and focus order are announced correctly
- Auditing a data visualisation component to confirm that chart data is available as a text alternative for screen reader users
- Reviewing a dynamic notification system to check that ARIA live regions announce updates without disrupting the user's current focus position
Use when
Test with a screen reader on every product that serves a public audience or must meet WCAG compliance — start testing early in the component-build phase, not as a final accessibility audit after development is complete.
Avoid when
Don't rely exclusively on automated accessibility checkers as a proxy for screen reader testing — automated tools catch structural issues but miss the lived experience problems that only emerge when navigating with audio output as your only feedback channel.
The fastest way to build empathy for screen reader users is to close your monitor, turn on VoiceOver, and attempt to complete your own product's core task — most designers discover within ninety seconds that the experience they shipped is unusable.
Real-world examples
- Twitter's early web interface had critical accessibility failures with screen readers — images had no alt text, icons had no labels — discovered only when blind journalist Sina Matagi published a detailed breakdown, prompting a major accessibility sprint.
- Apple's VoiceOver, first shipped in 2009, reads ARIA labels and roles from native UIKit components automatically, making apps built on iOS standard components accessible to 250 million VoiceOver users without extra engineering.
- GOV.UK Design System components are tested with JAWS, NVDA, and VoiceOver before release, ensuring any government service built with the system inherits screen reader compatibility without individual teams testing separately.