Performance Optimization

Tune VEXYL for maximum throughput and minimum latency.

Latency Optimization

To achieve sub-200ms latency, prioritize streaming providers and caching.

# Speed-optimized configuration
GATEWAY_MODE=false
STT_PROVIDER=deepgram
TTS_PROVIDER=deepgram
TTS_CACHE_ENABLED=true
VAD_REDEMPTION_FRAMES=6

Memory Management

For high-volume deployments, configure buffer limits to prevent memory exhaustion.

MAX_AUDIO_BUFFER_SIZE=50
MAX_PLAYBACK_QUEUE_SIZE=25
SESSION_TTL=1800

Scaling

VEXYL scales horizontally using Redis for session management. Use a load balancer (HAProxy/Nginx) to distribute traffic across multiple VEXYL instances.

REDIS_HOST=redis.internal
REDIS_PORT=6379