Post Content
Multimodal agents demand low latency, network resilience, clean audio, and turn detection all at once. In this session Jesse Hall (Developer Advocate, LiveKit) walks through building a real-time voice agent end-to-end, using Azure’s STT, LLM, and TTS models with LiveKit’s real-time infrastructure. Get an in-depth look into how the stack fits together and watch the completed agent answer questions about Microsoft Build.
𝗦𝗽𝗲𝗮𝗸𝗲𝗿𝘀:
* Jesse Hall
𝗦𝗲𝘀𝘀𝗶𝗼𝗻 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻:
This is one of many sessions from the Microsoft Build 2026 event. View even more sessions on-demand and learn about Microsoft Build at https://build.microsoft.com
ODSP937 | English (US) | Agents & apps
Pre-recorded | (300) Advanced
#MSBuild
Chapters:
0:00 – Overview of Topics: LiveKit, Cascaded Pipeline, Azure Integration
00:00:53 – Voice as a Real-Time Modality
00:01:52 – Scalability Requirements for Real-Time Applications
00:02:36 – Introduction to Agent’s SDK and Supported Languages
00:03:10 – Voice Activity Detection and Cost Optimization
00:03:48 – Zoomed-Out View: Global User Connectivity via WebRTC
00:06:41 – Introduction to Agent.py – core logic for the assistant
00:07:27 – Setting agent behavior and greeting users upon session start
00:08:23 – Configuring LiveKit and Azure environment variables including OpenAI API key Read More Microsoft Developer
