Zoom RTMS Transcript Service

A real-time transcript streaming service for Zoom meetings using the Zoom Real-Time Meeting Service (RTMS) API. This demo showcases live transcription capabilities and is designed to integrate with Groq Compound for advanced AI processing.

🚀 Features

• Real-Time Transcription: Live transcript streaming from Zoom meetings
• WebSocket Integration: Direct connection to Zoom's RTMS WebSocket endpoints
• Server-Sent Events: Real-time UI updates for transcript display
• Webhook Handling: Automatic handling of Zoom RTMS lifecycle events
• Security Headers: OWASP-compliant security headers implementation
• Cross-Platform: Runs on Deno, compatible with Val Town deployment
• Future AI Integration: Planned integration with Groq Compound for intelligent transcript processing

📋 Prerequisites

• Deno runtime
• Zoom App Marketplace account with RTMS enabled
• Environment variables configured (see below)

🔧 Setup

1. Clone the repository

   git clone <repository-url>
   cd zoom-rtms

2. Install dependencies

   # Dependencies are automatically managed by Deno
   # No npm install required

3. Configure environment variables

   Create a .env file in the project root:

   # Zoom RTMS Credentials (from Zoom App Marketplace)
   ZOOM_CLIENT_ID=your_zoom_client_id
   ZOOM_CLIENT_SECRET=your_zoom_client_secret
   ZOOM_SECRET_TOKEN=your_zoom_secret_token
   
   # Optional: Custom webhook path (defaults to /webhook)
   WEBHOOK_PATH=/webhook
   
   # Groq API Key
   GROQ_API_KEY=your_groq_api_key
   
   # Salesforce MCP Configuration
   SALESFORCE_MCP_URL=your_salesforce_mcp_url
   
   # AI Model Configuration (Optional - Override to test different models)
   # All default to their specified fallback models if not set
   
   # Intelligent routing decisions (default: openai/gpt-oss-20b)
   MODEL_ROUTER=openai/gpt-oss-20b
   
   # Discovery mode analysis (default: openai/gpt-oss-20b)
   MODEL_DISCOVERY=openai/gpt-oss-20b
   
   # Fact extraction from research (default: openai/gpt-oss-20b)
   MODEL_EXTRACTOR=openai/gpt-oss-20b
   
   # Text compression/distillation (default: openai/gpt-oss-20b)
   MODEL_COMPRESSOR=openai/gpt-oss-20b
   
   # Main inference with MCP tools (default: openai/gpt-oss-120b)
   MODEL_INFERENCE=openai/gpt-oss-120b
   
   # Direct answers without tools (default: openai/gpt-oss-120b)
   MODEL_DIRECT_ANSWER=openai/gpt-oss-120b
   
   # Multi-tool response synthesis (default: openai/gpt-oss-120b)
   MODEL_SYNTHESIS=openai/gpt-oss-120b

🏃 Running the Application

Development Mode

deno task print
# or directly:
deno serve --port 9995 --watch --allow-read --allow-env --allow-write --allow-net ./main.js

Production Deployment

deno task prod

The application will be available at:

• Main UI: http://localhost:9995/
• Webhook Endpoint: http://localhost:9995/webhook
• SSE Stream: http://localhost:9995/events

🏗️ Architecture

Components

1. Webhook Handler (/webhook)

   • Receives RTMS lifecycle events from Zoom
   • Handles meeting start/stop notifications
   • Validates webhook signatures

2. WebSocket Connections

   • Signaling WebSocket: Establishes RTMS session
   • Media WebSocket: Receives real-time transcript data

3. Server-Sent Events (/events)

   • Streams transcript data to connected clients
   • Cross-isolate broadcasting for Deno Deploy

4. Live UI (/)

   • Real-time transcript display
   • EventSource-based updates
   • Dark theme with monospace font (Menlo)

Data Flow

Zoom Meeting → RTMS Webhook → Signaling WS → Media WS → Transcripts → SSE → UI

🔐 Zoom RTMS Configuration

1. Create a Zoom App in the Zoom App Marketplace
2. Enable RTMS in your app settings
3. Configure Webhook URL pointing to your deployed endpoint
4. Set Event Types to include:
   • meeting.rtms_started
   • meeting.rtms_stopped
   • endpoint.url_validation

🌐 API Endpoints

Endpoint	Method	Description
/	GET	Live transcript viewer UI
/webhook	POST	Zoom RTMS webhook handler
/events	GET	Server-Sent Events stream

🚀 Future Integration: Groq Compound

This project is designed to integrate with Groq Compound for intelligent transcript processing:

• Real-time Analysis: Process transcripts as they're received
• Summarization: Generate meeting summaries automatically
• Action Items: Extract tasks and follow-ups
• Sentiment Analysis: Analyze participant engagement
• Translation: Real-time language translation
• Q&A: Answer questions about meeting content

Planned Features

• Live transcript enhancement
• Automated meeting minutes
• Keyword extraction and tagging
• Speaker identification improvements
• Content moderation and filtering

📚 Documentation

• Zoom RTMS Documentation
• Groq Compound API
• Hono Framework
• Deno Deploy

🔧 Development

Code Structure

├── main.js          # Main application
├── deno.json        # Deno configuration
├── deno.lock        # Dependency lock file
└── README.md        # This file

Key Technologies

• Runtime: Deno
• Framework: Hono
• WebSockets: Native WebSocket API
• Security: OWASP-compliant headers
• Deployment: Deno Deploy compatible

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request

📄 License

MIT License - see LICENSE file for details

⚠️ Disclaimer

This is a demonstration project for educational and development purposes. Ensure compliance with Zoom's terms of service and data privacy regulations when handling meeting transcripts. </xai:function_call">README.md