Skip to content

DEEPGRAM_SETUP.md

Latest commit

 

History

History
56 lines (42 loc) · 1.49 KB

DEEPGRAM_SETUP.md

File metadata and controls

56 lines (42 loc) · 1.49 KB

Deepgram Nova-3 Setup Guide

🚀 Quick Setup

1. Get Your Deepgram API Key

  1. Go to Deepgram Console
  2. Sign up for a free account (includes $200 in free credits)
  3. Navigate to "API Keys" in the dashboard
  4. Create a new API key and copy it

2. Set Up Environment Variables

Create a .env file in your project root:

# .env file
DEEPGRAM_API_KEY=your_deepgram_api_key_here

3. Install Dependencies

pip install -r requirements.txt

4. Run the Application

python assistant.py

🎯 Features

  • Nova-3 Model: Latest Deepgram speech-to-text model
  • Real-time Transcription: Live audio streaming
  • High Accuracy: Optimized for conversational AI
  • Smart Formatting: Automatic punctuation and capitalization

🔧 Configuration Options

The current setup uses:

  • Model: nova-2 (you can change to nova-3 when available)
  • Language: English (US)
  • Sample Rate: 16kHz
  • Channels: Mono
  • Format: Linear16

💡 Usage Tips

  1. Clear Speech: Speak clearly and at normal pace
  2. Quiet Environment: Minimize background noise
  3. Microphone: Use a good quality microphone for best results
  4. Commands: Say commands like "Question", "When am I free today?", etc.

🔍 Troubleshooting

  • API Key Issues: Make sure your .env file is in the project root
  • Audio Issues: Check microphone permissions
  • Connection Issues: Verify internet connection and API key validity