Features
Real-time streaming TTS with sub-200ms end-to-end latency
Coda model with LLM backbone and sub-100ms latency
Mist v3 model optimized for enterprise scale
Pronunciation correction without retraining (SpeechQA)
Flag low-confidence words before deployment
Multilingual support: English, Spanish, French, Portuguese, German, Japanese
Word-level timestamps for text-audio alignment
On-premises, VPC, or cloud API deployment
SOC 2 and HIPAA compliance
Supports barge-in and mid-utterance control
API-based TTS with multiple output formats (e.g., MP3)
CLI tool for batch processing
Real-time text-to-speech (Sonic-3.5)
Streaming speech-to-text (Ink-2)
State-space model architecture
Multi-region cloud API endpoints
On-premise (VPC) deployment
Line platform for building voice agents
Instant voice cloning (Pro+)
Professional voice cloning (Startup+)
Voice agent slot with concurrent calls
In-region data processing for compliance
Prepaid agent minutes with overage billing
