Playground
Generate speech from text using any model
Benchmark
Run all models on the same text and compare
Voice Cloning
Upload a reference voice (5–30 s) and clone it
Drop a WAV/MP3 file here
or click to browse
Tips for best results
- Use 5–20 seconds of clean audio
- Single speaker, minimal background noise
- Providing a transcript improves accuracy
- F5-TTS, SILMA, and Coqui XTTS-v2 give the strongest cloning options
Leaderboard
User ratings from this session
🇸🇦 Arabic Models
| Model | Ratings | Avg Rating | Avg Gen (s) | Avg RTF |
|---|
🇬🇧 English Models
| Model | Ratings | Avg Rating | Avg Gen (s) | Avg RTF |
|---|