AssemblyAI ships TTS-1: text-to-speech at $15 per 1M characters

AssemblyAI released AssemblyAI-TTS-1, its first text-to-speech model, aimed at voice agents on live phone calls. It costs $15 per 1M characters with no annual contract or concurrency tiers, and offers 4,000+ voices, instant cloning and six languages.

• Priced at $15 per 1M characters with unlimited concurrency
• 93.7% entity accuracy vs ElevenLabs v4's 93.6% in AssemblyAI's tests
• 159 ms median streaming latency on Cekura's benchmark, 8th of 19 models
• No SSML, speed control, pronunciation dictionaries or emotion tags yet