Speech-to-Text Input

Blindeforbundet and HLF both identified speech-to-text as a significant accessibility and efficiency improvement. For users with visual impairments or motor difficulties, typing on a mobile keyboard is slow and error-prone; dictation dramatically reduces the effort of completing a registration. For all users, the ability to narrate a summary immediately after a home visit - while the details are fresh - improves the quality and completeness of records. Blindeforbundet explicitly noted that recording during visits is unwanted and would inhibit open conversation, so the post-activity-only design constraint is a trust and safety requirement, not merely a technical choice.

The Speech Input Widget integrates with Flutter's speech_to_text plugin, which wraps the native iOS Speech framework and Android SpeechRecognizer. Microphone permission is requested lazily on first use with a clear purpose string. Transcription runs on-device where supported; no audio is sent to external servers. The Speech-to-Text Service manages the recognition session lifecycle (start, interim results, final result, error) and emits BLoC events consumed by the active form widget. The microphone button meets WCAG 2.2 AA touch target requirements and has an accessible label. An active recording indicator (visual + haptic) is displayed during dictation. If the device language is set to a supported Sami language, the service falls back to Norwegian recognition with a user-visible notice.