🎙️ Zero-Shot Voice Cloning
Provide a reference sample of any voice (5–10 seconds recommended). The MOSS neural audio tokenizer extracts vocal timbre and prosody to clone the voice instantly.
Live Microphone
Upload Audio File
Saved Voice Clone
00:00
Click the mic button to record · Speak clearly for 5 to 10 seconds
Step 2: Text for Cloned Voice to Speak
0 / 500
🗣️ Preset Neural Voices
Choose from high-fidelity pre-trained neural voices including English, Chinese, and Japanese speakers.
Text to Synthesize
0 / 500
📚 Custom Voice Library
Your saved custom speaker profiles. Use any profile instantly for zero-shot text-to-speech.