Clone Your Voice with AI: Local Setup with OpenVoice

What is OpenVoice?

OpenVoice is an open-source voice cloning model that can replicate a speaker's voice from a short audio sample (as little as 30 seconds). It supports emotion control, accent adjustment, and multilingual synthesis. All processing runs locally on your hardware.

When OpenVoice is configured for local-only operation and network integrations are disabled, its processing stays on the machine. Protect the resulting biometric data with appropriate local disk and access controls.

Requirements

• GPU: 6 GB+ VRAM recommended (CPU works but is 5–10x slower)
• Storage: ~2 GB for model weights
• Audio sample: 30–60 seconds of clean speech (minimal background noise)

Step 1: Install OpenVoice

Open Thumper-Run, find OpenVoice in the App Catalog, and click Install. Thumper downloads the model weights via the model pack system and configures the Python environment.

Step 2: Record or Upload a Voice Sample

For the best results:

• Record in a quiet room with minimal echo
• Speak naturally at a consistent volume
• Include varied intonation (questions, statements, emphasis)
• Use a decent microphone (even a phone mic works if you are close to it)

Save the recording as WAV or MP3 and load it into OpenVoice's reference speaker input.

Step 3: Generate Speech

Type or paste the text you want spoken in the cloned voice. Click Generate. On a mid-range GPU, a 30-second clip generates in under 10 seconds.

OpenVoice processes in two stages:

1. Base TTS generates speech in a neutral voice
2. Tone converter applies the reference speaker's voice characteristics

Tuning the Output

• Emotion: Adjust between neutral, happy, sad, angry, and whispering
• Speed: Control speaking rate without pitch distortion
• Language: Generate in English, Chinese, Japanese, Korean, and more—even if the reference sample is in a different language

Use Cases

• Audiobook narration with a consistent custom voice
• Podcast production – clone your own voice for quick re-records
• Accessibility – create a personal TTS voice for assistive technology
• Game development – prototype character voices before hiring voice actors

Privacy

Voice biometrics are sensitive data. With OpenVoice running locally through Thumper-Run, your voice samples and generated audio never leave your device. See our data privacy post for the full security model.

Ready to try it? Download Thumper-Run free →