Speech-to-Text Input

Both Blindeforbundet and HLF identified speech-to-text as a meaningful accessibility and efficiency improvement for peer mentors who find typing burdensome, have motor impairments, or regularly produce longer narrative reports. Blindeforbundet explicitly stated that recording during a home visit is unwanted - dictation is for the reporting phase afterward. For peer mentors logging detailed home-visit reports (Blindeforbundet's formalized structure) or complex activity summaries, voice input can reduce the time required to complete a report by an order of magnitude. This directly addresses the underreporting problem by lowering the effort barrier.

Use the speech_to_text Flutter plugin which wraps AVSpeechRecognizer (iOS) and SpeechRecognizer (Android). SpeechPermissionHandler manages microphone permission with a rationale dialog explaining dictation-only use (no live recording during visits). SpeechToTextService abstracts the plugin interface and emits a stream of interim and final transcription results to the SpeechInputWidget. The widget shows a visual recording indicator with an accessible announcement for screen reader users. Transcription results are editable before commit. The feature is implemented as a composable widget that can be attached to any AppTextField in the app. Background recording is explicitly disabled.