Text-to-Speech Providers
Select the best voice for your brand. VEXYL integrates with top-tier TTS engines for natural, low-latency synthesis.
Provider Comparison
| Provider | Latency | Quality | Best For |
|---|---|---|---|
| Deepgram | <200ms | Good | Speed, Cost |
| ElevenLabs | 300-800ms | Premium | Emotion, Quality |
| Azure | 200-400ms | Excellent | Styles (Happy, Sad) |
| Sarvam | 500ms | Good | Indian Accents |
Configuration
Deepgram (Recommended for Speed)
DEEPGRAM_API_KEY=your_key
TTS_PROVIDER=deepgram
DEEPGRAM_MODEL=aura-asteria-enElevenLabs (Recommended for Quality)
ELEVENLABS_API_KEY=your_key
TTS_PROVIDER=elevenlabs
ELEVENLABS_VOICE_ID=21m00Tcm4TlvDq8ikWAMTTS Caching
We strongly recommend enabling TTS caching to reduce latency to near-zero for common phrases and save up to 95% on API costs.
TTS_CACHE_ENABLED=true
TTS_CACHE_MAX_SIZE_MB=5000