[Shopping List] Task 5 – Conversational TTS Feedback #69
Description
Epic
Goal
Deliver reliable, natural-sounding spoken confirmations for every shopping list interaction, ensuring chefs receive immediate feedback without looking at the screen.
Deliverables
- Integration with OpenAI TTS pipeline (backend orchestration + caching strategy)
- Expo audio playback manager with queueing, interruption handling, and volume controls
- UX rules for when to speak confirmations vs. errors vs. prompts
- Offline/latency fallback messaging in UI
- Automated tests covering playback state machine and API error paths
Acceptance Criteria
- Each successful speech-driven action plays back confirmation within acceptable latency thresholds
- Errors surface with spoken guidance and visual alerts simultaneously
- Repeated prompts use cached audio when possible to minimize costs
- Audio subsystem coexists with ongoing speech capture without echo/feedback issues
Notes
- Coordinate with Task 4 to align intent responses with speech content
- Capture analytics on TTS failures for monitoring
- Ensure adherence to accessibility guidelines (option to mute or adjust volume)