Voice Cloner
macOS
Clone your own voice and have it speak texts or dialogs — entirely locally on your Mac. No account, no servers, no analytics; the only network connection is the one-time download of the speech model.
Last updated: September 7, 2026
1. Overview and Controller
What this is about
Voice Cloner is a Mac app that runs the VoxCPM2 speech model directly on the device: you record a short reference of your voice (or import an audio file), the app synthesizes text in that voice and combines several voices into dialogs. All computation happens on your Mac. Robware operates no servers for the app; there is no user account, no advertising, no telemetry, and no analytics services. Neither recordings nor texts nor generated audio files leave your computer.
Controller
The controller responsible for data processing is: Robware, owner Robin Wichmann, Hauptstr. 58c, 21640 Bliedersdorf, Germany. E-mail: kontakt@robware.de
Privacy contact
For privacy inquiries about this app, you can reach us at kontakt@robware.de.
2. Data Stored Locally
Reference voices
Microphone recordings or imported audio files are stored as WAV files together with a name and — optionally — the transcript of the spoken text. A person's voice is personal data; processing takes place exclusively on your device.
Syntheses and dialogs
Generated audio files, the entered text, the parameters used (e.g. language, style instruction, seed), and dialogs with their lines, speakers, and timing are stored in a local library so you can compare results and play them again.
Settings
Engine, synthesis, and audio preferences (such as loudness target, compute threads, playback options) are stored in the app's local preferences (UserDefaults).
Storage location
All data resides in the app's isolated container (App Sandbox) at ~/Library/Containers/de.robware.voicecloner. The app only accesses files you explicitly grant it via a file dialog for reading (import) or writing (export).
Legal basis
Processing is carried out to provide the app features you request, on the basis of Art. 6(1)(b) GDPR.
3. Microphone
The microphone is used only when you deliberately start a reference recording; macOS asks for your permission the first time. The recording ends when you click Stop or automatically after a short time and is saved locally as a WAV file. There is no background recording. You can revoke the permission at any time in System Settings under Privacy & Security → Microphone.
4. Model Download (the Only Network Connection)
On first launch (or at your request under Settings → Model → "Download again"), the app downloads the VoxCPM2 speech model (about 3 GB) from huggingface.co and verifies it with a SHA-256 checksum. In doing so, the platform operator, Hugging Face, Inc. (USA), receives the technically necessary connection data such as your IP address and the requested file; Hugging Face's privacy policy applies (https://huggingface.co/privacy). The app does not transmit any of your content — no recordings, no texts, no identifiers. Alternatively, you can install an existing model file via "Use existing file", in which case no network connection takes place at all. The legal basis is Art. 6(1)(b) GDPR (provision of the app feature).
5. What Voice Cloner Does Not Do
- No Robware servers: there is no backend infrastructure of ours that data could flow to.
- No analytics services, no crash reporting, no tracking, no ads.
- No account and no sign-up — after the model download, the app works entirely offline.
- No sharing with third parties: apart from the model download (Hugging Face) and obtaining the app via the Mac App Store (Apple), no service is involved.
6. Responsibility When Cloning Voices
Voice Cloner is made to clone your own voice. You may only use another person's voice as a reference if that person has expressly consented. You are responsible for how the generated audio is used: label AI-generated speech as such when you publish or share it, and do not use it for deception or impersonation. Robware has no access to your reference voices or results and can neither see nor influence their use.
7. Storage Period and Deletion
All data remains on your Mac until you delete it. Reference voices, syntheses, and dialogs can be removed directly in the app; the model can be deleted under Settings → Model. Deleting the app removes the entire container, including the library, audio files, and model. Since we have no access to your data, we cannot delete it for you.
8. Mac App Store and Components Used
The app is obtained through Apple's Mac App Store; Apple's privacy policy applies to purchase, download, and updates. The app contains the VoxCPM2 speech model (OpenBMB, Apache-2.0) and the inference library audio.cpp (Apache-2.0) with ggml (MIT). These components run exclusively locally and do not transmit any data.
Your Rights
Under the GDPR you have the right of access to the personal data concerning you (Art. 15), the right to rectification (Art. 16), erasure (Art. 17), restriction of processing (Art. 18), data portability (Art. 20), and the right to object to processing based on Art. 6(1)(f) GDPR (Art. 21). You may withdraw any consent you have given at any time with effect for the future. You also have the right to lodge a complaint with a data protection supervisory authority.
Contact
We answer questions about privacy in Voice Cloner at kontakt@robware.de.
