Compare all 6 OpenAI TTS voices side by side. Real neural TTS — no browser quirks.
Live: https://mvp.trollefsen.com/2026-04-14-voice-palette/
Voice Palette generates speech in all 6 OpenAI TTS voices simultaneously, so you can compare them directly on any input text. Pick a sample or type your own — then hit Generate All to hear how each voice handles it.
| Voice | Persona | Character |
|---|---|---|
| Alloy | The Narrator | Neutral, clear, versatile |
| Echo | The Broadcaster | Warm, resonant, professional |
| Fable | The Storyteller | Expressive, British warmth |
| Onyx | The Authority | Deep, powerful, commanding |
| Nova | The Guide | Bright, energetic, approachable |
| Shimmer | The Empath | Soft, expressive, nuanced |
- Generate All — fires all 6 voices in parallel, audio appears as each loads
- Per-voice generate — regenerate individual voices without losing the others
- Progress bar — live playback position per voice card
- Quality toggle — Standard (
tts-1) vs HD (tts-1-hd) - Sample texts — 4 presets covering narration, announcements, technical, conversational
- FastAPI backend — real OpenAI TTS, not browser speech synthesis
- React 19 + TypeScript
- Vite + Tailwind CSS v4
- FastAPI + uvicorn
- OpenAI TTS API (
tts-1/tts-1-hd)
git clone https://github.com/da-troll/voice-palette
cd voice-palette
# Backend
pip install fastapi uvicorn openai
# Set OPENAI_API_KEY or update backend/main.py
uvicorn backend.main:app --port 8000
# Frontend (dev mode, separate terminal)
npm install
npm run devWorks in all modern browsers. Audio playback uses standard HTML5 <audio> — no browser TTS quirks.
Built by the Nightly MVP Builder · 2026-04-14