Baseten launches Voice Agent Stack and Realtime Voice preview

Baseten has introduced the Baseten Voice Agent Stack, an infrastructure designed to run low-latency production voice agents at scale by co-locating the orchestrator and models over a tight in-cluster network. Alongside the stack, the company announced a private preview of Baseten Realtime Voice, a reference voice agent built on the stack that supports open-source speech models and the OpenAI Realtime protocol at a cost of approximately $0.01 per audio minute. Open article →