C
Codexa
Back to blog
Feature

How Live Caption Mode Works

Live Caption Mode is one of Codexa's most powerful features. Let's dive into how it works under the hood.

The Technology

We use Deepgram's real-time transcription API, which provides:

  • Sub-200ms latency
  • High accuracy for technical terms
  • Speaker diarization (coming soon)

    • How It Works

    1. **Audio Capture** - Codexa captures audio from your selected input device

    2. **Streaming** - Audio is streamed to Deepgram via WebSocket

    3. **Transcription** - Deepgram returns text in real-time

    4. **Display** - Captions appear in an always-on-top overlay

    Customization Options

  • Font size and style
  • Number of visible lines
  • Auto-hide when silent
  • Transparency settings

    • Best Practices

  • Use a good quality microphone
  • Reduce background noise
  • Position the overlay where it doesn't obstruct important content
    • Codexa - AI-Powered Development Assistant