Voice Intervention: microphone → speech-to-text → Intervention

[zeevenn]

Parent

#39 — PRD: AI Radio — Host-driven Channel experience with chat timeline

What to build

Wire the microphone button on the Radio Screen. Tapping it records a voice clip, transcribes it via a speech-to-text API, and feeds the resulting text into the same Intervention flow as the text input.

Flow:

1. User taps the mic button — recording starts (button shows active/recording state)
2. User taps again (or silence is detected) — recording stops
3. Audio is sent to a speech-to-text provider (e.g. OpenAI Whisper API or platform ASR)
4. Transcribed text is displayed in the input bar (user can review before submitting, or auto-submit)
5. Same Intervention flow as Text Intervention: track request, mood change, free instruction #48 from step 2 onward

Request microphone permission before the first recording attempt. If permission is denied, show an explanatory message.

Acceptance criteria

• Mic button triggers recording; visual state changes while recording
• Audio is transcribed and the text appears in the input bar or is auto-submitted
• Successful voice input triggers the same Intervention flow as text input (user-turn, host-turn, Programme regeneration)
• Microphone permission is requested on first use; denial is handled gracefully

Blocked by

• Text Intervention: track request, mood change, free instruction #48 (Text Intervention — voice feeds into the same flow)