MiraTTS Voice Cloning

High-quality voice synthesis with 100x realtime speed using optimized LMDeploy

Reference Audio

Upload a clear audio sample (3-30 seconds, 16kHz+)

Record directly in your browser

Text Input

Example Usage
  1. Upload or record a reference audio (your target voice)
  2. Enter text you want to synthesize in that voice
  3. Click generate and wait for the result

Tips:

  • Use clear reference audio without background noise
  • Keep reference audio between 3-30 seconds
  • Shorter text generates faster