Aratako · Voice
The VoiceDesign variant takes a text description of a voice in place of a reference recording, so no one's voice is cloned at the input. The card raises a different issue: a described voice can end up resembling a real person, unintentionally, as a probabilistic artifact of the latent space. It also embeds SilentCipher invisible watermarks in its output.
Summarized from the linked source documents, last checked 25 August 2026. Every claim links to its source; anything unconfirmed is marked unconfirmed.
MIT, verified on the upstream repo and the model cards. Commercial use permitted with no threshold. The cards add an ethics section (no cloning a real person's voice without explicit consent, no deepfakes). That is stated intent, not a term of MIT.
MIT claims nothing over output, but two card-level facts matter in practice: the VoiceDesign and v4 models embed SilentCipher invisible watermarks in what they generate, and the card notes a text-described voice can resemble a real person unintentionally, as a probabilistic artifact. Neither is a licence term.
Correction to the shipped catalog: The shipped catalog says "Unknown". It is MIT, per the upstream repo and the model card.
Based on the licence text as of the date below. Licences change; the links carry the authoritative version.
Transcribed from the shipped catalog, audiocpp_models_catalog.json — that file is authoritative if a number here disagrees with it.
audiocpp-irodori-tts-600m-v3-voicedesignirodori_ttsThere is no per-model install. One native binary runs all of these, so getting a model is a download of its weights and nothing else — no virtualenv, no torch pin, no CUDA wheel to match against your driver.
soundbox model install audiocpp
soundbox pull audiocpp-irodori-tts-600m-v3-voicedesign<models dir>/audiocpp/audiocpp-irodori-tts-600m-v3-voicedesign/ rather than in the shared Hugging Face cache, so a download stays scoped to the model you asked for.Primary sources: the model's own repository, the project it came from, and the conversion audio.cpp actually downloads. Never invented — an entry with nothing findable says so instead.
Licence details on these pages are summaries of the linked source documents, last checked 25 August 2026. They are not legal advice, and Soundbox is not affiliated with any model's authors — weights download from Hugging Face under the user's own account and terms. Company and model names identify whose work it is, and belong to their owners. The linked licence text is authoritative; commercial use should be checked against it, per model, before anything ships.