ClapVoice user guide

Start with the first line.

Set up your local models and get back to creating. This guide walks you through your first session.

Get started

ClapVoice currently supports Windows 10 and 11 (x64). Downloading models requires an internet connection and sufficient disk space. Performance depends on your hardware and the model size.

  1. Visit the download page for the latest Windows release status.
  2. Open the desktop app and select Models in the lower-left corner.
  3. Install the voice, transcription, or translation models you need.
  4. Open a workspace, add your material, and start creating.
You can use the desktop app without a website account. Your website account manages your profile, and desktop projects stay on your device.

Install local models

Speech generation and voice cloning

In Models → TTS, install the OmniVoice runtime and model, or select an existing weights folder and Python installation. Python 3.10–3.14 is required, with 3.11 or 3.12 recommended. The app creates its own virtual environment and starts the service for you.

Speech recognition

Under ASR, download a Whisper model or select an existing GGML .bin file. Transcription, live recording, and video dubbing use the same model.

Translation

Under Translation, install the Argos packages for your source and target languages to translate subtitles.

Create voices

In Voice Cloning, add a reference recording you have permission to use and enter its exact transcript. Import or paste your new script, select a language and quality, then generate, preview, and export as WAV. Press Ctrl + Enter to generate the current script.

In Voice Design, describe the voice in words and adjust the starting point and random seed. Preview the result, then save it to your voice library.

Use recordings you own or have explicit permission to use. Results vary with sample quality, language, and hardware.

Transcription

Import an audio or video file, or record in the app. Review and edit the transcript on the timeline, then copy the text or export to TXT, SRT, VTT, or DOCX. Recordings are kept if transcription fails so you can retry.

Video dubbing

Import a video or SRT subtitles, translate the dialogue, and edit it segment by segment. Assign voices, adjust timing, and preview before exporting WAV, MP4, or subtitles. Use the batch queue for multiple videos or languages.

Audiobooks

Import TXT or Markdown, organize chapters, and assign character voices and pauses. Resume generation where you left off or process multiple books in a batch. Export WAV, MP3, or chaptered M4B.

Website account

With an existing account, sign in, then open your account settings to change your display name or password. Email signup and recovery are currently unavailable.

Voices, scripts, and projects stay on your device. Cloud sync is currently unavailable.