Microsoft Foundry Agent Service is now generally available
for Voice Live integration, allowing developers to connect real-time
speech-to-text and text-to-speech to a Foundry Agent without writing their own
audio pipeline. Existing agent capabilities — including tool calling,
knowledge, memory, guardrails, and enterprise integrations — continue to work
seamlessly and are now enhanced with low-latency speech interactions. The agent
provides a real-time streaming voice channel with interrupt handling, turn-taking,
barge-in support, connection management, natural voices, and avatars for
production phone and web experiences. For customers building customer service
agents, learning companions, and in-car assistants, this provides a single
supported path from Foundry agent logic to live audio input and output, instead
of requiring them to stitch together Speech, the agent runtime, and WebRTC on
their own.
Learn
more.