Voice Preview Generation

[lbedner]

Scope

Generate sample audio for voice preview functionality.

Tasks

1. Add preview endpoint

POST /ai/voice/preview
Body: {"voice": "nova", "text": "Hello, I'm Nova."}
Response: audio/mpeg

2. Pre-generate samples for each voice

• Standard intro: "Hello, I'm {voice_name}. How can I help you today?"
• Options:
  • Cache as static files (pre-generated)
  • Generate on-demand (first request)
  • Hybrid: pre-generate common voices, on-demand for others

3. Frontend integration

• Play preview when user clicks voice card
• Show loading indicator during generation
• Handle playback errors gracefully

Files

• app/components/backend/api/ai/router.py (add endpoint)
• frontend/dashboard/components/VoicePreview.tsx

Dependencies

• Requires Voice Catalog Infrastructure #480 (Voice Catalog Infrastructure)
• Integrates with Voice Settings Tab Frontend #482 (Voice Settings Tab Frontend)

Verification

curl -X POST http://localhost:8000/ai/voice/preview \
  -H "Content-Type: application/json" \
  -d '{"voice": "nova", "text": "Hello!"}' \
  --output preview.mp3