Skip to main content

Model Selection

Echophrase uses OpenAI’s Whisper models. Choose the right one for your needs.

Available Models

ModelSizeSpeedAccuracyBest For
Tiny~75 MBFastestGoodQuick notes, low-end hardware
Base~150 MBFastBetterDaily use, balanced
Small~500 MBMediumGreatProfessional use
Medium~1.5 GBSlowerExcellentHigh accuracy needs
Large~3 GBSlowestBestMaximum accuracy

Recommendations

  • CPU only: Start with Tiny or Base
  • NVIDIA GPU: Base or Small for best balance
  • High-end GPU: Medium or Large for maximum accuracy

Changing Models

  1. Open Echophrase settings
  2. Navigate to AI Model
  3. Select your preferred model
  4. Wait for the download to complete (~1-2 minutes)
Models are downloaded once and cached locally. Switching between downloaded models is instant.