Supported
- A purchased or otherwise legitimately obtained Supertonic 3 Voice Builder JSON.
- Local schema validation, local preview and browser-local synthesis.
- Re-exporting the canonical tensor-only backup.
Free local migration tool
Keep your Supertonic 3 voice after Voice Builder closes
Download the model-specific JSON you already purchased, import it into UnlimitedTTS, preview it privately and keep using the voice with the open-weight Supertonic 3 model.
Download deadline: Voice Builder says existing users can download purchased embeddings until .
Voice Builder has closed. You can still import a Supertonic 3 JSON you downloaded earlier, but UnlimitedTTS cannot recover a file that was not backed up before shutdown.
Local preview needs about 380 MB
Supertonic 3 runs entirely on your device. The model downloads once and is cached by this browser for later use.
The model may fail on iPhone, iPad and other low-memory devices. A desktop browser is more reliable.
Three steps to preserve a Voice Builder voice
The import is a migration of an existing purchased embedding—not new voice cloning. Everything happens on your device.
-
Download the Supertonic 3 JSON
Sign in to Voice Builder My Page and choose the file explicitly labeled Supertonic 3 JSON.
Find the separately labeled Supertonic 3 JSON in the independent backup you retained before shutdown.
-
Import and preview locally
Choose the JSON, confirm its version and your usage rights, enter a display name, then complete a short browser-local preview before saving.
-
Keep an independent backup
Retain the original download. After import, open the voice manager and choose Download backup so browser storage is never your only copy.
What UnlimitedTTS imports—and what it does not
Not supported
- Supertonic 2 JSON as a supported drop-in v3 style. “v2-compatible” refers to the ONNX inference interface, not the voice embedding.
- Recovering Play/API voices or files that were never downloaded.
- Creating a new voice from reference audio.
Private by default, portable by design
UnlimitedTTS reads only the two style tensors needed for Supertonic 3. Source metadata is discarded, and the filename, display name, preview text and tensor values are never sent to analytics.
What stays local
The imported tensor data and your display name are saved in IndexedDB for this browser profile. Speech generation runs through Supertonic 3 using WebGPU or WebAssembly on your device.
What analytics records
Only aggregate funnel stages, normalized error codes, migration phase and whether storage is browser-local. No voice identity, filename, vector, checksum or text content is included.
Official references
Check the source notices before the deadline. UnlimitedTTS is an independent project and is not endorsed by or affiliated with Supertone.
Supertonic Voice Builder migration FAQ
Which Supertonic file should I download before Voice Builder closes?
Download the file explicitly labeled Supertonic 3 JSON from Voice Builder My Page. Supertonic 2 and 3 exports are model-specific even though their visible JSON structure can look the same.
Can I import a Voice Builder JSON after August 31, 2026?
Yes, if you already downloaded and retained the Supertonic 3 JSON. UnlimitedTTS imports an existing backup locally; it cannot recover a file that was never downloaded before Voice Builder closed.
Can my browser delete an imported voice?
Yes. Imported voices currently live in browser storage, not an UnlimitedTTS account. Clearing site data, private-browsing cleanup, a browser profile reset, storage pressure, or occasionally a browser update can remove that data. Keep the original JSON and a separate backup.
Does the imported voice file leave my device?
No in the current local mode. Validation, preview, storage and Supertonic 3 speech generation happen in your browser. UnlimitedTTS analytics never receives the filename, voice name, metadata, tensor values or checksum.
Are Supertonic 2 voice files compatible with Supertonic 3?
Not as an officially supported drop-in voice style. Supertone's v2-compatible statement refers to the public ONNX inference interface, while its Voice Builder documentation describes separate, version-specific JSON files. A v2 file can have the same shape and pass structural validation, but that does not prove its embedding represents the same voice in v3. Supertone has not published a direct v2-to-v3 embedding conversion.
Can I export an imported voice from UnlimitedTTS again?
Yes. Open the importer, find the saved voice and choose Download backup. The exported file contains the canonical voice tensors without the source metadata that may have appeared in the original file.
Preserve the file while you still can
Download the Supertonic 3 JSON before the Voice Builder portal closes, then import it here and keep a second backup.
If you retained a Supertonic 3 JSON backup, you can still import, preview and use it locally.
Last verified August 11, 2026. Product names and trademarks belong to their respective owners. This page provides engineering and migration guidance, not legal advice.