Story
Cartesia
DailyDaily Bots Achieves Human-Conversational Voice AI Speed with Cartesia Sonic
Daily, a voice and video infrastructure provider, accelerated real-time conversational AI by implementing Cartesia's Sonic text-to-speech model as the primary voice provider for Daily Bots.
Value results
| Category | Value result |
|---|---|
| Cost | **180ms** time-to-first-byte for voice inference including network overhead |
| Speed | Voice-to-voice latencies as fast as **500ms** (human conversational speed) |
| Scale | Natural voice support across **13 languages** with customizable voice options |
| Capability | Hundreds of developer sign-ups on Day 1 of launch, reflecting pent-up demand for hosted voice AI agents |
| Capability | Unlocks new enterprise use cases including customer support, appointment scheduling, and virtual persona interactions |
Daily is a leading voice and video infrastructure provider serving thousands of developers globally with low-latency SDKs and a global edge network spanning 75 points of presence. Earlier this year, Daily launched Daily Bots, an open-source cloud platform enabling developers to build adaptive voice AI agents using any language model (LLM). To deliver the conversational speed and voice quality required for enterprise voice AI applications, Daily selected Cartesia's Sonic as its primary text-to-speech provider for the platform.
Daily Bots optimized adaptive voice AI for exceptional conversational experience and enterprise reliability by integrating Cartesia Sonic, which leverages state-of-the-art State Space Models (SSMs) research to achieve unprecedented performance in real-time voice applications. Sonic enables natural, responsive voice interactions across 13 languages with support for customized voices, while Daily's global edge network ensures reliable delivery to over 5 billion users worldwide. The integration addresses developer demand for hosted voice AI agents built on open standards including RTVI and Pipecat frameworks.
The partnership delivered measurable improvements in voice AI performance and developer adoption. Daily Bots achieves voice-to-voice latencies as fast as 500ms—human conversational speed—with Cartesia consistently delivering a time-to-first-byte for voice inference of 180ms or better, including network overhead. The platform now powers use cases including customer support, appointment scheduling, and virtual persona interactions. Hundreds of developers signed up on the first day of Daily Bots launch, demonstrating strong market demand for enterprise-grade voice AI infrastructure.
