AssemblyAI launches first TTS model with 90ms first-audio latency

AssemblyAI has released its first text-to-speech model, AssemblyAI-TTS-1, built for real-time voice agents and available on LiveKit Inference; trained on real phone conversations, it delivers first audio in just 90ms.

2026-11-03 ~ 2026-11-04 · 2 related posts

• AssemblyAI launches TTS-1, its first text-to-speech model built for voice agents — AssemblyAI · 2026-11-03
• AssemblyAI launches first TTS model on LiveKit: phone-call-trained, 90ms first audio, $15/1M chars — AssemblyAI · 2026-11-04