| title: XTTS Voice Studio | |
| emoji: 🎙️ | |
| colorFrom: yellow | |
| colorTo: yellow | |
| sdk: docker | |
| app_port: 7860 | |
| pinned: false | |
| # XTTS v2 Voice Studio | |
| A multilingual text-to-speech studio powered by **Coqui XTTS v2**, served via **FastAPI** (no Gradio). | |
| ## Features | |
| - 16 supported languages (Arabic, English, French, …) | |
| - Voice cloning from uploaded audio samples | |
| - Persistent voice library | |
| - Generation history with playback & download | |
| - Fully custom React UI | |
| ## Usage | |
| 1. Upload a reference audio clip (WAV / MP3 / FLAC, ≥ 6 s recommended). | |
| 2. Type your text and pick a language. | |
| 3. Adjust advanced parameters if needed. | |
| 4. Click **⚡ توليد الصوت** and wait — CPU inference takes ~30–90 s per request. | |
| ## Notes | |
| - Running on **CPU**; generation is slower than GPU but fully functional. | |
| - The XTTS v2 model (~1.8 GB) is downloaded on first startup and cached. | |
| - `COQUI_TOS_AGREED=1` is set automatically — by using this Space you agree to the [Coqui TTS terms](https://coqui.ai/cpml). | |