AssemblyAI-TTS-1 is AssemblyAI's first text-to-speech model, providing natural speech for voice agents with 90 ms streaming latency at $15 per million characters. It features leading accuracy on independent benchmarks, instant voice cloning, over 4,000 voices, unlimited concurrency without contracts, and is available today.
AssemblyAI-TTS-1 is AssemblyAI’s text-to-speech model for conversational voice agents. It turns text into streaming audio through WebSocket and HTTP interfaces, with an OpenAI-compatible route also available. The product supports English, Spanish, French, German, Italian, and Portuguese, and offers a voice library and voice cloning. Additional features include word-level timestamps, interruption handling, and US or EU processing options. It can be used in an existing voice-agent stack or alongside AssemblyAI speech-to-text services.
