Text-to-speech — natural AI voice audio from any text (Kokoro). Pay-per-call USDC, no API key. For voice agents, narration, and accessibility — cheaper than dedicated TTS providers.