Boson AI · Voice

Higgs Audio v3 TTS 4B

Commercial use: conditions applyBoson Higgs TTS 3 Research and Non-Commercial License

The default is research and non-commercial: deploying it in production, behind a hosted API, or inside a product needs a commercial licence from Boson AI. A Creator Use Grant on top of that lets individual creators produce creative content, including monetised podcasts, video and social posts, for free, provided they credit Boson AI's Higgs Audio in the audio or prominently in accompanying text. Check which of the two cases your use falls under.

The licence

Summarized from the linked source documents, last checked 25 August 2026. Every claim links to its source; anything unconfirmed is marked unconfirmed.

Boson Higgs TTS 3 Research and Non-Commercial License

Commercial use: conditions apply

Research and non-commercial by default. A separate commercial licence from Boson AI is required to deploy it in production, offer it as a hosted API, embed it in a product, or resell it. The Creator Use Grant is the exception: monetised creative content is permitted for free with attribution.

Conditions

  • Product, API, hosted-service and resale use require a separate commercial licence from Boson AI.
  • Creator Use Grant: monetised creative content is allowed if you credit Boson AI's Higgs Audio in the audio itself or prominently in accompanying text.
  • The acceptable-use terms prohibit voice cloning without consent, impersonation, fraud, election deception and biometric surveillance.

What about the audio you generate?

Where legally required, AI-generated audio must be disclosed. The Creator Use Grant's attribution requirement is itself an obligation attached to what you publish, so here the output carries a duty as well as the weights.

Correction to the shipped catalog: The shipped catalog says "Unknown". It is Boson AI's research and non-commercial licence with a Creator Use Grant. The names differ: the model is "Higgs Audio v3 TTS 4B" and the licence is named for "Higgs TTS 3".

Based on the licence text as of the date below. Licences change; the links carry the authoritative version.

Specs

Transcribed from the shipped catalog, audiocpp_models_catalog.json — that file is authoritative if a number here disagrees with it.

Vendor
Boson AI
Catalog id
audiocpp-higgs-audio-tts
audio.cpp loader family
higgs_audio_tts
Can do
Text to speech
Weights file
Higgs-Audio-v3-TTS-4B-GGUF/higgs-audio-v3-tts-4b-q8_0.gguf
Account needed to download
No
Max clip length
60s
Sample rate
44.1kHz mono
Approx. VRAM
3.0 GB
In the picker by default
No — opt in from Settings
Run end-to-end here
No — catalogued, not tested here

Getting it

There 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.

From the CLI
soundbox model install audiocpp
soundbox pull audiocpp-higgs-audio-tts
  • The binary is fetched from audio.cpp's GitHub releases the first time you ask for it and cached under your own data directory. Soundbox ships no native binaries and vendors none.
  • Weights land in <models dir>/audiocpp/audiocpp-higgs-audio-tts/ rather than in the shared Hugging Face cache, so a download stays scoped to the model you asked for.
  • This one is opt-in: switch it on in the Models page's “Enable more models” section before it appears in the picker.

Worth knowing

  • Opt-in and untested here.

Sources

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.