zoom-agent

Overview ¶

Example: Voice agent in Zoom meetings

This example demonstrates how to add a voice agent to Zoom meetings. Due to Zoom SDK complexity, we recommend using Recall.ai as the bot infrastructure.

Architecture with Recall.ai:

┌──────────────────────────────────────────────────────────────────┐
│                        Zoom Meeting                               │
│   ┌─────────┐  ┌─────────┐  ┌─────────────────────────────────┐  │
│   │  User 1 │  │  User 2 │  │     Recall.ai Bot               │  │
│   └─────────┘  └─────────┘  └────────────────┬────────────────┘  │
└──────────────────────────────────────────────┼────────────────────┘
                                               │
                                   WebSocket   │ (Audio stream)
                                               ▼
                              ┌─────────────────────────────────┐
                              │        OmniVoice Agent          │
                              │  ┌─────┐ ┌─────┐ ┌─────┐        │
                              │  │ STT │→│ LLM │→│ TTS │        │
                              │  └─────┘ └─────┘ └─────┘        │
                              └─────────────────────────────────┘

Why Recall.ai instead of native Zoom SDK: - Zoom doesn't provide a simple API for real-time audio - Native SDK requires one process per meeting, complex scaling - Recall.ai handles Zoom SDK complexity, provides WebSocket stream - You can be live in hours instead of months