Text to Speech API - Up To 40% Faster Globally

ElevenLabs has introduced multi-region serving for their Text to Speech API, significantly improving performance by automatically routing requests to the nearest backend in the US, Netherlands, or Singapore, thus enhancing the time to first byte (TTFB) without requiring any code changes. This update, leveraging upgraded GPUs and an optimized inference stack called Flash v2.5, results in a 20-40% reduction in perceived latency for international developers, with measured TTFB improvements across various global locations ranging from 50-200ms. The enhancements are particularly beneficial for voice agents and real-time applications, as they contribute to more natural conversations and consistent user experiences worldwide. Developers using the api.elevenlabs.io endpoint are already benefiting from this global routing, while those preferring US server usage can opt out by using a specific base URL. Enterprise customers with regional data residency needs are advised to contact sales for further assistance.