LlamaIndex, a leading data framework for large language model (LLM) applications, is set to host a technical workshop on August 14, 2025, demonstrating how to build real-time AI agents capable of processing live voice data from Zoom meetings. The workshop, featuring @ojusave and @tuanacelik, will provide a comprehensive blueprint for LLM orchestration with live voice data, leveraging Zoom's newly introduced Real-Time Media Streams (RTMS) functionality.
Zoom RTMS offers developers secure, real-time access to audio, video, and transcript data directly from Zoom Meetings via secure WebSocket. This innovation transforms live meeting content into structured data streams, enabling AI-driven transcription, compliance monitoring, summarization, and workflow automation. Arun Janakiraman, Group Product Manager of Apps at Zoom, emphasized that these innovations reflect Zoom's commitment to equipping developers with tools for the future of work and communication.
LlamaIndex's framework is designed to help developers build LLM applications by providing a comprehensive toolkit for data augmentation. Its AI agents are described as LLM-powered knowledge workers that can dynamically ingest and modify data from various tools, going beyond static data pipelines. This capability is crucial for creating intelligent, event-driven systems that can summarize conversations, detect intent, and generate action items from live audio.
The workshop will cover practical steps, including setting up Zoom RTMS to capture live audio and using transcript chunks as LLM context. Participants will learn to build intelligent, event-driven agents that can perform tasks like creating meeting notes. This integration addresses the growing demand for AI solutions that can provide immediate insights and automate tasks within collaborative environments.
The collaboration between LlamaIndex and Zoom RTMS highlights a significant trend in the AI and communication technology sectors: the move towards more responsive and integrated AI applications. Such advancements are poised to enhance productivity across various industries, from financial services requiring real-time compliance monitoring to sales teams needing instant customer interaction analysis. The workshop aims to empower developers to create production-grade AI systems that seamlessly interact with streaming audio data.