AssemblyAI launches TTS-1 speech API at $15 per million characters

AssemblyAI has fully opened access to AssemblyAI-TTS-1, its first text-to-speech model, with pricing set at $15 per million characters. The API is aimed at voice agents used in real-time phone call scenarios and does not require annual contracts or the purchase of concurrency tiers. According to the release, the model was trained on millions of hours of audio data, including real conversational data, and uses continuous audio representations to improve realism while reducing time to first audio output. The launch also includes a library of more than 4,000 voices and an instant voice cloning feature that requires only a short recording and confirmation of speaker consent. At launch, the model supports six languages: English, Spanish, French, German, Italian, and Portuguese. Developers can integrate it through an HTTP API, WebSocket API, or an OpenAI-compatible route. The update was cited by Techub News, with MarkTechPost named as the source.