Fabio/voice_audio_multi
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1
voice_audio_multi/docs/model_licenses.md
Fabio c36510c21e
Some checks failed
Linux Build / Linux cpu (push) Has been cancelled
Details
Linux Build / Linux cpu-clang (push) Has been cancelled
Details
Linux Build / Linux vulkan (push) Has been cancelled
Details
mac-build / macOS CPU (push) Has been cancelled
Details
Nix Build / Nix (cpu) (push) Has been cancelled
Details
Nix Build / Nix (vulkan) (push) Has been cancelled
Details
windows-build / Windows CPU (push) Has been cancelled
Details
Initial import: audio.cpp with Orange Pi documentation
2026-11-04 19:05:22 +01:00

184 lines
34 KiB
Markdown
Raw Blame History

# Model Licenses
audio.cpp is licensed under Apache-2.0. That license does not cover the model
weights it loads: each model keeps the license of its original release, and a GGUF
conversion keeps the license of the model it was converted from.
This page lists, per model family, where the original weights are published, their
license, and whether that license allows commercial use. The license text at the source
is what counts. This page is a summary to help you find it, not legal advice.
Licenses change. If a row is wrong or out of date, please open a PR.
## Commercial use
| Value | Meaning |
|---|---|
| Yes | The license allows commercial use. Attribution, notice, share-alike or use-policy terms may apply. |
| Conditional | Commercial use is allowed only under a condition, such as a revenue or user limit, a territory, registration, or a notice in the product. Notes names the condition. |
| No | Research or non-commercial use only. |
| Unclear | The sources conflict, or no license is stated. Notes says why. |
## Families
| Family | Original weights | License | Commercial use | Notes | Checked |
|---|---|---|---|---|---|
| `ace_step` | [ACE-Step/Ace-Step1.5](https://huggingface.co/ACE-Step/Ace-Step1.5) | MIT | Yes | | 2026-09-21 |
| `apollo` | [JusperLee/Apollo](https://huggingface.co/JusperLee/Apollo) | CC-BY-SA-4.0 | Yes | | 2026-09-21 |
| `ast_audioset` | [MIT/ast-finetuned-audioset-10-10-0.4593](https://huggingface.co/MIT/ast-finetuned-audioset-10-10-0.4593) | BSD-3-Clause | Yes | | 2026-09-26 |
| `audio8_asr` | [Edge0/Audio8-ASR-0.1B](https://huggingface.co/Edge0/Audio8-ASR-0.1B) | CC-BY-NC-4.0 | No | | 2026-09-21 |
| `audio8_tts` | [Edge0/Audio8-TTS-Preview-0.6b](https://huggingface.co/Edge0/Audio8-TTS-Preview-0.6b) | Apache-2.0 | Yes | | 2026-09-21 |
| `audio_flamingo` | [nvidia/audio-flamingo-3-hf](https://huggingface.co/nvidia/audio-flamingo-3-hf), [nvidia/audio-flamingo-next-hf](https://huggingface.co/nvidia/audio-flamingo-next-hf), [nvidia/music-flamingo-2601-hf](https://huggingface.co/nvidia/music-flamingo-2601-hf) | NVIDIA OneWay Noncommercial | No | Upstream model-card licenses and terms apply. SpotSound adapter is MIT; base-model terms still apply. | 2026-10-26 |
| `audiosr` | [haoheliu/audiosr_basic](https://huggingface.co/haoheliu/audiosr_basic) | Apache-2.0 | Yes | | 2026-09-21 |
| `auk` | [tencent/AuK](https://huggingface.co/tencent/AuK)<br>[tencent/AuK-Flash](https://huggingface.co/tencent/AuK-Flash) | MIT | Yes | | 2026-09-21 |
| `breeze_tts` | [BreezeBlue/Breeze-TTS-2](https://huggingface.co/BreezeBlue/Breeze-TTS-2) | [BreezeBlue Research and Non-Commercial License](https://huggingface.co/BreezeBlue/Breeze-TTS-2/blob/main/LICENSE) | No | | 2026-09-21 |
| `bs_roformer` | Not linked | Apache-2.0, as listed in the audio.cpp GGUF repo | Unclear | The source of the ep368 checkpoint is not documented. | 2026-09-21 |
| `builtin_audio_utils` | [xiph/rnnoise](https://github.com/xiph/rnnoise)<br>[Rikorose/DeepFilterNet](https://github.com/Rikorose/DeepFilterNet) | BSD-3-Clause (RNNoise); MIT or Apache-2.0 (DeepFilterNet) | Yes | Weights are downloaded separately. | 2026-09-21 |
| `canary_asr` | [nvidia/canary-180m-flash](https://huggingface.co/nvidia/canary-180m-flash) | CC-BY-4.0 | Yes | | 2026-09-21 |
| `ced` | [mispeech/ced-tiny](https://huggingface.co/mispeech/ced-tiny), [ced-mini](https://huggingface.co/mispeech/ced-mini), [ced-small](https://huggingface.co/mispeech/ced-small), [ced-base](https://huggingface.co/mispeech/ced-base) | Apache-2.0 | Yes | Checkpoint license; the separate upstream training repository is GPL-3.0. | 2026-10-09 |
| `chatterbox` | [ResembleAI/chatterbox](https://huggingface.co/ResembleAI/chatterbox) | MIT | Yes | | 2026-09-21 |
| `chatterbox_turbo` | [ResembleAI/chatterbox-turbo](https://huggingface.co/ResembleAI/chatterbox-turbo) | MIT | Yes | | 2026-09-21 |
| `citrinet_asr` | [nvidia/stt_en_citrinet_256_ls](https://huggingface.co/nvidia/stt_en_citrinet_256_ls) | CC-BY-4.0 | Yes | | 2026-09-21 |
| `cohere_asr` | [CohereLabs/cohere-transcribe-03-2026](https://huggingface.co/CohereLabs/cohere-transcribe-03-2026) | Apache-2.0 | Yes | | 2026-09-21 |
| `confucius4_r2t2` | [netease-youdao/Confucius4-R2T2](https://huggingface.co/netease-youdao/Confucius4-R2T2) | [NetEase Youdao Model License](https://github.com/netease-youdao/Confucius4-R2T2/blob/master/MODEL_LICENSE) | Conditional | A separate license is needed above RMB 1 billion yearly revenue or 100 million monthly users. The model and its outputs may not be used to improve other commercial AI models. | 2026-09-21 |
| `confucius4_tts` | [netease-youdao/Confucius4-TTS](https://huggingface.co/netease-youdao/Confucius4-TTS) | Apache-2.0 | Yes | | 2026-09-21 |
| `controlfoley` | [YJX-Xiaomi/ControlFoley](https://huggingface.co/YJX-Xiaomi/ControlFoley) | CC-BY-NC-4.0 | No | | 2026-09-21 |
| `cosyvoice3` | [FunAudioLLM/Fun-CosyVoice3-0.5B-2512](https://huggingface.co/FunAudioLLM/Fun-CosyVoice3-0.5B-2512) | Apache-2.0 | Yes | | 2026-09-21 |
| `crisperwhisper` | [nyralabs/CrisperWhisper2.0_large](https://huggingface.co/nyralabs/CrisperWhisper2.0_large) | nyra health Non-Commercial Research License | No | The upstream license also restricts generated outputs; inference software is MIT. | 2026-09-29 |
| `dots_tts` | [dots-studio/dots.tts-soar](https://huggingface.co/dots-studio/dots.tts-soar)<br>[dots-studio/dots.tts-mf](https://huggingface.co/dots-studio/dots.tts-mf)<br>[dots-studio/dots.tts.edit](https://huggingface.co/dots-studio/dots.tts.edit) | Apache-2.0 | Yes | The dots.tts.edit card defers to the [dots.tts project](https://github.com/studio-dots-ai/dots.tts), which is Apache-2.0. | 2026-09-21 |
| `dramabox` | [ResembleAI/Dramabox](https://huggingface.co/ResembleAI/Dramabox) | [LTX-2 Community License](https://huggingface.co/ResembleAI/Dramabox/blob/main/LICENSE) | Conditional | A paid license is needed from USD 10M yearly revenue. Generated content must be disclosed as machine-generated. Not for products that compete with Lightricks'. The Gemma 3 text encoder falls under the Gemma Terms of Use. | 2026-09-21 |
| `earshot_vad` | [pykeio/earshot](https://github.com/pykeio/earshot) | MIT OR Apache-2.0 | Yes | Bundled in `assets/framework/models`; converted from the 1.2.2 crate. | 2026-10-26 |
| `echo_tts` | [jordand/echo-tts-base](https://huggingface.co/jordand/echo-tts-base)<br>[jordand/fish-s1-dac-min](https://huggingface.co/jordand/fish-s1-dac-min) | CC-BY-NC-SA-4.0 | No | | 2026-09-21 |
| `f5_tts` | [SWivid/F5-TTS](https://huggingface.co/SWivid/F5-TTS)<br>[SWivid/Habibi-TTS](https://huggingface.co/SWivid/Habibi-TTS) | CC-BY-NC-4.0 (F5-TTS); CC-BY-NC-SA-4.0 (Habibi-TTS) | No | | 2026-09-21 |
| `firered_audio` | [FireRedTeam/FireRedAudio](https://huggingface.co/FireRedTeam/FireRedAudio) | Apache-2.0 | Yes | | 2026-09-21 |
| `firered_vad` | [FireRedTeam/FireRedVAD](https://huggingface.co/FireRedTeam/FireRedVAD) | Apache-2.0 | Yes | VAD and Stream-VAD checkpoints. | 2026-10-07 |
| `fireredtts3` | [FireRedTeam/FireRedTTS3](https://huggingface.co/FireRedTeam/FireRedTTS3) | Apache-2.0 | Yes | | 2026-09-21 |
| `fish_audio` | [fishaudio/s2-pro](https://huggingface.co/fishaudio/s2-pro) | [Fish Audio Research License](https://huggingface.co/fishaudio/s2-pro/blob/main/LICENSE.md) | No | Commercial use needs a separate license from Fish Audio. | 2026-09-21 |
| `fish_drama3` | [fishaudio/drama-3](https://huggingface.co/fishaudio/drama-3) | [Fish Audio Community License v1.0](https://huggingface.co/fishaudio/drama-3/blob/main/LICENSE) | Conditional | Community commercial use has a revenue threshold; hosted services and other uses may require a separate license. See the source terms. | 2026-10-14 |
| `fun_asr_nano` | [FunAudioLLM/Fun-ASR-Nano-2512](https://huggingface.co/FunAudioLLM/Fun-ASR-Nano-2512) | Apache-2.0 | Yes | | 2026-09-21 |
| `gigaam_asr` | [ai-sage/GigaAM-v3](https://huggingface.co/ai-sage/GigaAM-v3), [ai-sage/GigaAM-Multilingual](https://huggingface.co/ai-sage/GigaAM-Multilingual) | MIT | Yes | | 2026-09-25 |
| `glm_tts` | [zai-org/GLM-TTS](https://huggingface.co/zai-org/GLM-TTS) | MIT | Yes | | 2026-09-21 |
| `gradium_phonon` | [gradium/phonon-31m-en](https://huggingface.co/gradium/phonon-31m-en)<br>[gradium/phonon-31m-en-audiocpp](https://huggingface.co/gradium/phonon-31m-en-audiocpp) | Apache-2.0 | Yes | Both packages come from Gradium's repositories: the Q8_0 GGUF and the published safetensors checkpoint, each with its default voice. The 40m checkpoints ([gradium/phonon-40m-en](https://huggingface.co/gradium/phonon-40m-en) and other languages) are CC-BY-NC-4.0 and gated, and are not packaged. | 2026-10-21 |
| `granite5asr` | [ibm-granite/granite-speech-5.0-470m-turboctc](https://huggingface.co/ibm-granite/granite-speech-5.0-470m-turboctc) | Apache-2.0 | Yes | | 2026-09-21 |
| `haas_asr` | [third-intelligence/Haas-ja-v1.0](https://huggingface.co/third-intelligence/Haas-ja-v1.0/tree/660728d39019e5e065046154beca44ac9a66a540) | Apache-2.0 | Yes | Japanese ASR variant with a Kimi Delta Attention predictor. Upstream states that both code and model weights are Apache-2.0. | 2026-11-03 |
| `heartmula` | [HeartMuLa/HeartMuLa-oss-3B](https://huggingface.co/HeartMuLa/HeartMuLa-oss-3B)<br>[HeartMuLa/HeartCodec-oss-20260123](https://huggingface.co/HeartMuLa/HeartCodec-oss-20260123) | Apache-2.0 | Yes | | 2026-09-21 |
| `higgs_audio_stt` | [bosonai/higgs-audio-v3-stt](https://huggingface.co/bosonai/higgs-audio-v3-stt) | Apache-2.0 | Yes | | 2026-09-21 |
| `higgs_audio_tts` | [bosonai/higgs-tts-3-4b](https://huggingface.co/bosonai/higgs-tts-3-4b) | [Boson Higgs TTS 3 Research and Non-Commercial License](https://huggingface.co/bosonai/higgs-tts-3-4b/blob/main/LICENSE) | No | | 2026-09-21 |
| `htdemucs` | [facebookresearch/demucs](https://github.com/facebookresearch/demucs) | MIT | Yes | | 2026-09-21 |
| `htdemucs_6stems` | [facebookresearch/demucs `5c90dfd2`](https://dl.fbaipublicfiles.com/demucs/hybrid_transformer/5c90dfd2-34c22ccb.th) | MIT | Yes | | 2026-09-26 |
| `hviske_asr` | [syvai/hviske-v5.3](https://huggingface.co/syvai/hviske-v5.3) | CC-BY-NC-4.0 | No | | 2026-09-21 |
| `hviske_asr_v6` | [syvai/hviske-v6](https://huggingface.co/syvai/hviske-v6) | CC-BY-NC-4.0 | No | Retain upstream NOTICE and third-party license notices. | 2026-10-07 |
| `index_echo` | [IndexTeam/Index-Echo-S2TT-2B](https://huggingface.co/IndexTeam/Index-Echo-S2TT-2B)<br>[IndexTeam/Index-Echo-S2TT-9B](https://huggingface.co/IndexTeam/Index-Echo-S2TT-9B) | Apache-2.0 | Yes | S2TT checkpoints only. | 2026-10-01 |
| `index_tts2` | [IndexTeam/IndexTTS-2](https://huggingface.co/IndexTeam/IndexTTS-2)<br>[IndexTeam/IndexTTS-2.5](https://huggingface.co/IndexTeam/IndexTTS-2.5) | [bilibili Model Use License](https://huggingface.co/IndexTeam/IndexTTS-2.5/blob/main/LICENSE) | Conditional | A separate license is needed above RMB 1 billion yearly revenue or 100 million monthly users. The model and its outputs may not be used to improve other commercial AI models. | 2026-09-21 |
| `inflect_v2` | [owensong/Inflect-Micro-v2](https://huggingface.co/owensong/Inflect-Micro-v2)<br>[owensong/Inflect-Nano-v2](https://huggingface.co/owensong/Inflect-Nano-v2) | Apache-2.0 | Yes | | 2026-09-21 |
| `irodori_tts` | [Aratako/Irodori-TTS-500M-v3](https://huggingface.co/Aratako/Irodori-TTS-500M-v3)<br>[Aratako/Irodori-TTS-600M-v3-VoiceDesign](https://huggingface.co/Aratako/Irodori-TTS-600M-v3-VoiceDesign)<br>[Aratako/Irodori-TTS-v4.1-Small](https://huggingface.co/Aratako/Irodori-TTS-v4.1-Small) | MIT | Yes | | 2026-09-21 |
| `kitten_tts` | [KittenML/kitten-tts-mini-0.8](https://huggingface.co/KittenML/kitten-tts-mini-0.8) | Apache-2.0 | Yes | | 2026-09-21 |
| `kitten_tts2` | [KittenML/kitten-tts-2](https://huggingface.co/KittenML/kitten-tts-2) | [Stellon Labs Community License](https://huggingface.co/KittenML/kitten-tts-2/blob/main/LICENSE.md) | Conditional | Commercial registration and revenue/funding limits; distribution and product attribution requirements apply. Decoder components retain their upstream terms. | 2026-10-02 |
| `kokoro_tts` | [hexgrad/Kokoro-82M](https://huggingface.co/hexgrad/Kokoro-82M) | Apache-2.0 | Yes | | 2026-09-21 |
| `kroko_asr` | [Banafo/Kroko-ASR](https://huggingface.co/Banafo/Kroko-ASR) | CC-BY-SA | Yes | Stated in the model card text only: the LICENSE file is empty and no version is named. Applies to the community models; Kroko's commercial models are licensed separately. | 2026-09-21 |
| `kugelaudio` | [kugelaudio/kugelaudio-0-open](https://huggingface.co/kugelaudio/kugelaudio-0-open) | MIT | Yes | Includes the upstream preset voice features. | 2026-09-28 |
| `lfm2_audio` | [LiquidAI/LFM2.5-Audio-1.5B](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B)<br>[LiquidAI/LFM2.5-Audio-1.5B-JP](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B-JP) | [LFM Open License v1.0](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B/blob/main/LICENSE) | Conditional | Commercial use is licensed only to organizations with less than USD 10M yearly revenue. The audio encoder is based on canary-180m-flash (CC-BY-4.0). | 2026-09-25 |
| `lfm3_audio` | [LiquidAI/LFM3-Audio-2B](https://huggingface.co/LiquidAI/LFM3-Audio-2B)<br>[LiquidAI/LFM3-Audio-2B-GGUF](https://huggingface.co/LiquidAI/LFM3-Audio-2B-GGUF) | [LFM Open License v1.0](https://huggingface.co/LiquidAI/LFM3-Audio-2B/blob/main/LICENSE) | Conditional | Commercial use is licensed only to organizations with less than USD 10M yearly revenue. | 2026-10-27 |
| `latentsync` | [ByteDance/LatentSync-1.6](https://huggingface.co/ByteDance/LatentSync-1.6)<br>[InsightFace buffalo_l](https://github.com/deepinsight/insightface#license) | Apache-2.0 (LatentSync); non-commercial research (InsightFace weights) | No | Combined package includes face detection and landmark weights with separate restrictions. | 2026-09-30 |
| `liveavatar` | [Quark-Vision/Live-Avatar](https://huggingface.co/Quark-Vision/Live-Avatar)<br>[Wan-AI/Wan2.2-S2V-14B](https://huggingface.co/Wan-AI/Wan2.2-S2V-14B) | Apache-2.0 | Yes | | 2026-09-21 |
| `magenta_realtime_2` | [google/magenta-realtime-2](https://huggingface.co/google/magenta-realtime-2) | CC-BY-4.0 | Yes | Retain attribution and the upstream usage terms. | 2026-10-27 |
| `magpie_tts` | [nvidia/magpie_tts_multilingual_357m](https://huggingface.co/nvidia/magpie_tts_multilingual_357m) | [NVIDIA Open Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/) | Yes | | 2026-09-21 |
| `marblenet_vad` | Not linked | Not stated | Unclear | Bundled in `assets/framework/models`; the checkpoint's source is not documented. | 2026-09-21 |
| `maya1` | [maya-research/maya1](https://huggingface.co/maya-research/maya1)<br>[hubertsiuzdak/snac_24khz](https://huggingface.co/hubertsiuzdak/snac_24khz) | Apache-2.0 (Maya1); MIT (SNAC) | Yes | The combined GGUF contains weights from both projects. | 2026-09-25 |
| `meanvc2` | [ASLP-lab/MeanVC2](https://huggingface.co/ASLP-lab/MeanVC2) | Apache-2.0 | Yes | | 2026-09-21 |
| `mel_band_roformer` | [mlx-community/mel-roformer-mlx](https://huggingface.co/mlx-community/mel-roformer-mlx) | MIT | Yes | | 2026-09-21 |
| `micro_wake_word` | [ESPHome Okay Nabu](https://github.com/esphome/micro-wake-word-models/blob/main/models/v2/okay_nabu.json) | Apache-2.0 | Yes | Default checkpoint. Upstream [license scope](https://github.com/esphome/micro-wake-word-models#license) explicitly includes model weights and manifests; retain LICENSE and [NOTICE](https://github.com/esphome/micro-wake-word-models/blob/main/NOTICE). Other user-supplied models may have different terms. | 2026-10-14 |
| `midashenglm_gen` | [mispeech/midashenglm-gen](https://huggingface.co/mispeech/midashenglm-gen) | Apache-2.0 | Yes | | 2026-09-21 |
| `minimax_h3` | [MiniMaxAI/MiniMax-H3](https://huggingface.co/MiniMaxAI/MiniMax-H3) | [MiniMax H3 Community License](https://huggingface.co/MiniMaxAI/MiniMax-H3/blob/main/LICENSE) | Conditional | Grants no rights in the EU, the UK, South Korea or the USA. Elsewhere, written authorization is needed above USD 20M yearly revenue, and "MiniMax H3" must be shown in the product's user interface. | 2026-09-21 |
| `minimax_music3` | [MiniMaxAI/MiniMax-Music3](https://huggingface.co/MiniMaxAI/MiniMax-Music3) | [MiniMax-Music3 Community License](https://huggingface.co/MiniMaxAI/MiniMax-Music3/blob/main/LICENSE) | Conditional | "MiniMax-Music3" must be shown in the product's user interface. Written authorization is needed above USD 20M yearly revenue. Content published in public must be disclosed as machine-generated. | 2026-09-21 |
| `miocodec` | [Aratako/MioCodec-25Hz-44.1kHz-v2](https://huggingface.co/Aratako/MioCodec-25Hz-44.1kHz-v2) | MIT | Yes | | 2026-09-21 |
| `ming3_audio` | [inclusionAI/Ming-3.0-audio-tiny](https://huggingface.co/inclusionAI/Ming-3.0-audio-tiny) | MIT | Yes | Includes the Thinker, Talker, audio VAE, and CAMPPlus speaker encoder. Retain the upstream MIT notice. | 2026-10-21 |
| `ming3_tts` | [inclusionAI/Ming-3.0-tts-0.5B](https://huggingface.co/inclusionAI/Ming-3.0-tts-0.5B) | MIT | Yes | Retain the inclusionAI MIT notice; the official package includes its audio VAE and CAMPPlus speaker encoder. | 2026-10-20 |
| `miotts` | [Aratako/MioTTS-1.7B](https://huggingface.co/Aratako/MioTTS-1.7B) | Apache-2.0 | Yes | | 2026-09-21 |
| `mira_tts` | [YatharthS/MiraTTS](https://huggingface.co/YatharthS/MiraTTS) | CC-BY-NC-SA-4.0 | No | | 2026-09-21 |
| `mms_forced_aligner` | [MahmoudAshraf/mms-300m-1130-forced-aligner](https://huggingface.co/MahmoudAshraf/mms-300m-1130-forced-aligner) | CC-BY-NC-4.0 | No | | 2026-09-21 |
| `moonshine_asr` | [moonshine-ai/moonshine-streaming-tiny](https://huggingface.co/moonshine-ai/moonshine-streaming-tiny)<br>[moonshine-ai/moonshine-streaming-small](https://huggingface.co/moonshine-ai/moonshine-streaming-small)<br>[moonshine-ai/moonshine-streaming-medium](https://huggingface.co/moonshine-ai/moonshine-streaming-medium) | MIT | Yes | | 2026-09-21 |
| `moss_transcribe_diarize` | [OpenMOSS-Team/MOSS-Transcribe-Diarize](https://huggingface.co/OpenMOSS-Team/MOSS-Transcribe-Diarize) | Apache-2.0 | Yes | | 2026-09-21 |
| `moss_tts_local` | [OpenMOSS-Team/MOSS-TTS-Local-Transformer-v1.5](https://huggingface.co/OpenMOSS-Team/MOSS-TTS-Local-Transformer-v1.5) | Apache-2.0 | Yes | | 2026-09-21 |
| `moss_tts_nano` | [OpenMOSS-Team/MOSS-TTS-Nano-100M](https://huggingface.co/OpenMOSS-Team/MOSS-TTS-Nano-100M) | Apache-2.0 | Yes | | 2026-09-21 |
| `moss_tts_v15` | [OpenMOSS-Team/MOSS-TTS-v1.5](https://huggingface.co/OpenMOSS-Team/MOSS-TTS-v1.5) | Apache-2.0 | Yes | | 2026-09-22 |
| `moss_ttsd` | [OpenMOSS-Team/MOSS-TTSD-v1.0](https://huggingface.co/OpenMOSS-Team/MOSS-TTSD-v1.0) | Apache-2.0 | Yes | License declared on the official model card; bundled [MOSS-Audio-Tokenizer](https://huggingface.co/OpenMOSS-Team/MOSS-Audio-Tokenizer#license) is also Apache-2.0. | 2026-10-13 |
| `moss_voicegen` | [OpenMOSS-Team/MOSS-VoiceGenerator](https://huggingface.co/OpenMOSS-Team/MOSS-VoiceGenerator) | Apache-2.0 | Yes | | 2026-09-21 |
| `mossformer2` | [alibabasglab/MossFormer2_SS_16K](https://huggingface.co/alibabasglab/MossFormer2_SS_16K) | Apache-2.0 | Yes | Two-speaker separation checkpoint from ClearerVoice-Studio. | 2026-10-04 |
| `muscriptor` | [MuScriptor/muscriptor-small](https://huggingface.co/MuScriptor/muscriptor-small) | CC-BY-NC-4.0 | No | | 2026-09-21 |
| `nemotron_3_diar` | [nvidia/Nemotron-3-Diarization](https://huggingface.co/nvidia/Nemotron-3-Diarization) | [OpenMDW-1.1](https://openmdw.ai/license/1-1/) | Yes | | 2026-09-24 |
| `nemotron_asr` | [nvidia/nemotron-3.5-asr-streaming-0.6b](https://huggingface.co/nvidia/nemotron-3.5-asr-streaming-0.6b) | [OpenMDW-1.1](https://openmdw.ai/license/1-1/) | Yes | | 2026-09-21 |
| `nemotron_transcribe` | [nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B](https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B) | [OpenMDW-1.1](https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B/blob/a99e33f698257b8b69ca7834f19c73582eb1fa7e/LICENSE) | Yes | Preserve the checkpoint and component notices in converted packages. | 2026-10-29 |
| `neutts` | [neuphonic/neutts-2e](https://huggingface.co/neuphonic/neutts-2e) | [NeuTTS Open License v1.0](https://huggingface.co/neuphonic/neutts-2e/blob/main/LICENSE) | Conditional | Commercial use is licensed only while yearly revenue stays under USD 5M. | 2026-09-21 |
| `niagara_asr` | [abr-ai/niagara-19m-batch.en](https://huggingface.co/abr-ai/niagara-19m-batch.en)<br>[abr-ai/niagara-38m-batch.en](https://huggingface.co/abr-ai/niagara-38m-batch.en) | [Applied Brain Research Open License v1.1](https://www.appliedbrainresearch.com/license) | Conditional | Any use while yearly gross revenue, including affiliates, stays under USD 1M; above that, non-commercial research only. | 2026-09-21 |
| `omnivoice` | [k2-fsa/OmniVoice](https://huggingface.co/k2-fsa/OmniVoice)<br>[mohammedaly22/VoiceTut-TTS](https://huggingface.co/mohammedaly22/VoiceTut-TTS) | [CC-BY-NC](https://huggingface.co/k2-fsa/OmniVoice#license) | No | Changed from Apache-2.0 on 2026-07-03, "due to constraints from its training data (e.g., Emilia)". The code is Apache-2.0. The audio tokenizer, [bosonai/higgs-audio-v2-tokenizer](https://huggingface.co/bosonai/higgs-audio-v2-tokenizer), is under the Boson Higgs Audio 2 Community License. VoiceTut-TTS is a fine-tune of OmniVoice that its model card declares as Apache-2.0; it reuses the OmniVoice audio tokenizer unchanged. | 2026-10-04 |
| `open_wake_word` | [OpenWakeWord Alexa v0.1](https://github.com/dscripka/openWakeWord/releases/tag/v0.5.1) | CC-BY-NC-SA-4.0 | No | Upstream [pretrained-model license](https://github.com/dscripka/openWakeWord#license); the Apache-2.0 code license does not cover these weights. | 2026-10-17 |
| `outetts` | [OuteAI/Llama-OuteTTS-1.0-1B](https://huggingface.co/OuteAI/Llama-OuteTTS-1.0-1B) | CC-BY-NC-SA-4.0 | No | | 2026-09-21 |
| `owsm` | [espnet/owsm_v4_base_102M](https://huggingface.co/espnet/owsm_v4_base_102M)<br>[espnet/owsm_v4_small_370M](https://huggingface.co/espnet/owsm_v4_small_370M)<br>[espnet/owsm_v4_medium_1B](https://huggingface.co/espnet/owsm_v4_medium_1B) | CC-BY-4.0 | Yes | Attribution required. | 2026-10-02 |
| `owsm_ctc` | [espnet/owsm_ctc_v4_1B](https://huggingface.co/espnet/owsm_ctc_v4_1B) | CC-BY-4.0 | Yes | Attribution required. | 2026-10-02 |
| `parakeet_tdt` | [nvidia/parakeet-tdt-0.6b-v3](https://huggingface.co/nvidia/parakeet-tdt-0.6b-v3) | CC-BY-4.0 | Yes | | 2026-09-21 |
| `parakeet_tdt` | [oruk/orukeet](https://huggingface.co/oruk/orukeet) | CC-BY-SA-4.0 | Yes | The Orukeet r3 weight variant. | 2026-09-21 |
| `phonon2_asr` | [FermionResearch/Phonon-2](https://huggingface.co/FermionResearch/Phonon-2) | CC-BY-4.0 | Yes | English-only weight variant; retain upstream NOTICE and attribution when converting/distributing. | 2026-10-09 |
| `personaplex` | [nvidia/personaplex-7b-v1](https://huggingface.co/nvidia/personaplex-7b-v1) | [NVIDIA Open Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/) | Yes | | 2026-09-21 |
| `piper_tts` | [rhasspy/piper-voices](https://huggingface.co/rhasspy/piper-voices/tree/main/en/en_US/lessac/medium) (`en_US-lessac-medium`) | MIT | Unclear | The voice's model card points to the [Lessac Blizzard 2013 dataset license](https://www.cstr.ed.ac.uk/projects/blizzard/2013/lessac_blizzard2013/), which allows non-commercial use only. | 2026-09-21 |
| `plapre` | [syvai/plapre-3](https://huggingface.co/syvai/plapre-3) | CC-BY-4.0 | Yes | Includes Kanade (MIT), WavLM and CosyVoice2 HiFT components; retain their upstream notices as well. | 2026-10-17 |
| `pocket_tts` | [kyutai/pocket-tts](https://huggingface.co/kyutai/pocket-tts), including [French](https://huggingface.co/kyutai/pocket-tts/tree/main/languages/french) | CC-BY-4.0 | Yes | Packaged language checkpoints, including the six-layer French model, share the upstream model-card license. Attribution required. | 2026-10-14 |
| `pocket_tts_small` | [kyutai/pocket-tts](https://huggingface.co/kyutai/pocket-tts) | CC-BY-4.0 | Yes | English 18M, 7M and 2M checkpoints, including ReFL variants. Attribution required. | 2026-11-03 |
| `pulsevad` | [AydinAdnan/PulseVAD](https://github.com/AydinAdnan/PulseVAD) | MIT | Yes | | 2026-09-21 |
| `qwen3_asr` | [Qwen/Qwen3-ASR-0.6B](https://huggingface.co/Qwen/Qwen3-ASR-0.6B)<br>[Qwen/Qwen3-ASR-1.7B-hf](https://huggingface.co/Qwen/Qwen3-ASR-1.7B-hf)<br>[mohammedaly22/QwenCleo-ASR](https://huggingface.co/mohammedaly22/QwenCleo-ASR) | Apache-2.0 | Yes | | 2026-10-04 |
| `qwen3_forced_aligner` | [Qwen/Qwen3-ForcedAligner-0.6B](https://huggingface.co/Qwen/Qwen3-ForcedAligner-0.6B) | Apache-2.0 | Yes | | 2026-09-21 |
| `qwen3_tts` | [Qwen/Qwen3-TTS-12Hz-0.6B-Base](https://huggingface.co/Qwen/Qwen3-TTS-12Hz-0.6B-Base)<br>[Qwen/Qwen3-TTS-12Hz-1.7B-Base](https://huggingface.co/Qwen/Qwen3-TTS-12Hz-1.7B-Base)<br>[Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice](https://huggingface.co/Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice)<br>[Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign](https://huggingface.co/Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign) | Apache-2.0 | Yes | | 2026-09-21 |
| `realtime_reuse` | [nvidia/Real-time_RE-USE](https://huggingface.co/nvidia/Real-time_RE-USE) | [NVIDIA One-Way Noncommercial License (NSCLv1)](https://github.com/NVlabs/HMAR/blob/main/LICENSE) | No | Causal checkpoint with exit-layer and lookahead controls. | 2026-11-02 |
| `reuse` | [nvidia/RE-USE](https://huggingface.co/nvidia/RE-USE) | [NVIDIA One-Way Noncommercial License (NSCLv1)](https://github.com/NVlabs/HMAR/blob/main/LICENSE) | No | Research and noncommercial use only. | 2026-09-27 |
| `rvc` | [lj1995/VoiceConversionWebUI](https://huggingface.co/lj1995/VoiceConversionWebUI) | MIT | Unclear | The base models (HuBERT, RMVPE) are MIT. The packaged voices (`manthos`, `chocola`, `fraise`) have no documented source or license. | 2026-09-21 |
| `sam_audio` | [facebook/sam-audio-small](https://huggingface.co/facebook/sam-audio-small) | [SAM License](https://huggingface.co/facebook/sam-audio-small/blob/main/LICENSE) | Yes | The packaged Google T5 Base text encoder retains its Apache-2.0 license. | 2026-09-26 |
| `samsone` | [SamsungLabs/samsone](https://github.com/SamsungLabs/samsone) | Not stated | Unclear | The upstream repository and v1.0.0 release do not publish model-weight license terms. | 2026-09-26 |
| `sanotts` | [ampixa/sanoTTS](https://huggingface.co/ampixa/sanoTTS) | GPL-3.0 | Yes | Copyleft. | 2026-09-21 |
| `seed_vc` | [Plachta/Seed-VC](https://huggingface.co/Plachta/Seed-VC)<br>[mlx-community/SeedVC-MLX](https://huggingface.co/mlx-community/SeedVC-MLX) | GPL-3.0 | Yes | Copyleft. | 2026-09-21 |
| `sense_asr` | [FunAudioLLM/SenseVoiceSmall](https://huggingface.co/FunAudioLLM/SenseVoiceSmall) | [FunASR Model Open Source License v1.1](https://github.com/modelscope/FunASR/blob/main/MODEL_LICENSE) | Yes | Credit the source and authors, and keep the model names. | 2026-09-21 |
| `sheetsage2` | [m-a-p/SheetSage2](https://huggingface.co/m-a-p/SheetSage2) | CC-BY-NC-4.0 | No | | 2026-09-21 |
| `sherpa_kws` | [Zipformer zh-en 3M, 2025-12-20](https://modelscope.cn/models/pkufool/icefall-kws-zipformer-zh-en-3M-2025-12-20) | Apache-2.0 | Yes | Default Chinese-English checkpoint; [model author confirmed the license](https://github.com/k2-fsa/sherpa-onnx/issues/3802#issuecomment-6052663364). Other user-supplied exports may have different terms. | 2026-10-14 |
| `sidon` | [sarulab-speech/sidon-v0.1](https://huggingface.co/sarulab-speech/sidon-v0.1) | MIT | Yes | | 2026-09-28 |
| `silero_vad` | [snakers4/silero-vad](https://github.com/snakers4/silero-vad) | MIT | Yes | Bundled in `assets/framework/models`. | 2026-09-21 |
| `smart_turn` | [pipecat-ai/smart-turn](https://github.com/pipecat-ai/smart-turn) | BSD-2-Clause | Yes | Smart Turn v3.2. | 2026-09-30 |
| `soprano_tts` | [ekwek/Soprano-1.1-80M](https://huggingface.co/ekwek/Soprano-1.1-80M) | Apache-2.0 | Yes | | 2026-09-21 |
| `sopro_tts` | [samuel-vitorino/sopro-v2-turbo](https://huggingface.co/samuel-vitorino/sopro-v2-turbo) | Apache-2.0 | Yes | | 2026-09-21 |
| `sortformer_diar` | [nvidia/diar_sortformer_4spk-v1](https://huggingface.co/nvidia/diar_sortformer_4spk-v1) | CC-BY-NC-4.0 | No | | 2026-09-21 |
| `sortformer_diar_v2` | [nvidia/diar_streaming_sortformer_4spk-v2.1](https://huggingface.co/nvidia/diar_streaming_sortformer_4spk-v2.1) | [NVIDIA Open Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/) | Yes | | 2026-09-21 |
| `stable_audio` | [stabilityai/stable-audio-3-small-sfx](https://huggingface.co/stabilityai/stable-audio-3-small-sfx)<br>[stabilityai/stable-audio-3-small-music](https://huggingface.co/stabilityai/stable-audio-3-small-music)<br>[stabilityai/stable-audio-3-medium](https://huggingface.co/stabilityai/stable-audio-3-medium) | [Stability AI Community License](https://huggingface.co/stabilityai/stable-audio-3-small-sfx/blob/main/LICENSE.md) | Conditional | Free while the yearly revenue of you and your affiliates stays under USD 1M; above that an enterprise license is needed. Register commercial use with Stability AI, and display "Powered by Stability AI". The Gemma text encoder falls under the Gemma Terms of Use. | 2026-09-21 |
| `supertonic` | [Supertone/supertonic-3](https://huggingface.co/Supertone/supertonic-3) | [BigScience OpenRAIL-M](https://huggingface.co/Supertone/supertonic-3/blob/main/LICENSE) | Yes | Its use restrictions apply and must be passed on. | 2026-09-21 |
| `tf_gridnet` | [ESPnet WSJ0-2mix checkpoint](https://drive.google.com/file/d/1TasZxZSnbSPsk_Wf7ZDhBAigS6zN8G9G/view) | Not stated | Unclear | The checkpoint archive does not specify a weight license; ESPnet code is Apache-2.0. | 2026-10-03 |
| `tone_color_vc` | [myshell-ai/OpenVoiceV2](https://huggingface.co/myshell-ai/OpenVoiceV2) | MIT | Yes | Standalone V2 converter. | 2026-09-27 |
| `underdog_bark3` | [ConwayResearch/Underdog-Bark-3.0-Pro](https://huggingface.co/ConwayResearch/Underdog-Bark-3.0-Pro) | CC-BY-4.0 | Yes | English-only ASR variant; converted GGUFs retain the upstream NOTICE, full license, and conversion provenance. Credit Conway Research, NVIDIA and Moondream. | 2026-10-29 |
| `underdog_bark3` | [ConwayResearch/Underdog-Bark-3.0-Plus](https://huggingface.co/ConwayResearch/Underdog-Bark-3.0-Plus) | CC-BY-4.0 | Yes | English-only compressed ASR checkpoint; native GGUF expansion retains NOTICE/license and conversion/rounding provenance. Credit Conway Research, NVIDIA and Moondream. | 2026-10-29 |
| `underdog_bark3` | [ConwayResearch/Underdog-Bark-3.0](https://huggingface.co/ConwayResearch/Underdog-Bark-3.0) | CC-BY-4.0 | Yes | Standard English-only compressed ASR checkpoint; converted GGUFs retain NOTICE/license and conversion provenance. Credit Conway Research, NVIDIA and Moondream. | 2026-10-29 |
| `universr` | [woongzip1/universr-audio](https://huggingface.co/woongzip1/universr-audio)<br>[woongzip1/universr-speech](https://huggingface.co/woongzip1/universr-speech) | CC-BY-4.0 | Yes | | 2026-09-21 |
| `vevo2` | [RMSnow/Vevo2](https://huggingface.co/RMSnow/Vevo2) | CC-BY-NC-ND-4.0 | No | | 2026-09-21 |
| `vibeasr` | [microsoft/VibeVoice-ASR-BitNet](https://huggingface.co/microsoft/VibeVoice-ASR-BitNet) | MIT | Yes | | 2026-09-21 |
| `vibevoice` | [microsoft/VibeVoice-1.5B](https://huggingface.co/microsoft/VibeVoice-1.5B)<br>[vibevoice/VibeVoice-7B](https://huggingface.co/vibevoice/VibeVoice-7B) | MIT | Yes | | 2026-09-21 |
| `vibevoice_asr` | [microsoft/VibeVoice-ASR](https://huggingface.co/microsoft/VibeVoice-ASR) | MIT | Yes | | 2026-09-21 |
| `vibevoice_asr_streaming` | [microsoft/VibeVoice-ASR-Streaming-7B](https://huggingface.co/microsoft/VibeVoice-ASR-Streaming-7B)<br>[microsoft/VibeVoice-ASR-Streaming-1.5B](https://huggingface.co/microsoft/VibeVoice-ASR-Streaming-1.5B) | MIT | Yes | | 2026-09-21 |
| `vietneu_tts` | [pnnbao-ump/VieNeu-TTS-v3-Turbo](https://huggingface.co/pnnbao-ump/VieNeu-TTS-v3-Turbo) | Apache-2.0 | Yes | | 2026-09-21 |
| `voxcpm1` | [openbmb/VoxCPM-0.5B](https://huggingface.co/openbmb/VoxCPM-0.5B) | Apache-2.0 | Yes | | 2026-09-21 |
| `voxcpm2` | [openbmb/VoxCPM2](https://huggingface.co/openbmb/VoxCPM2) | Apache-2.0 | Yes | | 2026-09-21 |
| `voxtral_realtime` | [mistralai/Voxtral-Mini-4B-Realtime-2602](https://huggingface.co/mistralai/Voxtral-Mini-4B-Realtime-2602) | Apache-2.0 | Yes | | 2026-09-21 |
| `voxtral_tts` | [mistralai/Voxtral-4B-TTS-2603](https://huggingface.co/mistralai/Voxtral-4B-TTS-2603) | CC-BY-NC-4.0 | No | Includes the upstream preset voice embeddings, which share the model's non-commercial license. | 2026-10-07 |
| `whistle_asr` | [Cactus-Compute/whistle](https://huggingface.co/Cactus-Compute/whistle/tree/b358ddadd89b7a713b5aa131f23032d3cca1b251) | Apache-2.0 | Yes | Pinned release includes the tokenizer, mel filterbank, and Hadamard permutations; the converter embeds its LICENSE. Independent asset provenance remains unverified. No reference code is bundled; see [provenance](../tools/community_models/whistle_asr/PROVENANCE.md). | 2026-10-06 |
| `yue2` | [m-a-p/YuE2-3B](https://huggingface.co/m-a-p/YuE2-3B) | CC-BY-NC-4.0 | No | | 2026-09-21 |
| `zipvoice` | [k2-fsa/ZipVoice](https://huggingface.co/k2-fsa/ZipVoice) | Not stated | Unclear | The model card states no license; the [code](https://github.com/k2-fsa/ZipVoice) is Apache-2.0. | 2026-09-21 |
## Adding or updating a row
- Keep one row per family, in alphabetical order. Split a family into several rows only
when its models carry different licenses.
- Link the original weights, not a GGUF conversion: the conversion keeps the original's
license.
- Read the license on the original model card and in its LICENSE file. Use the SPDX
identifier where one exists (`Apache-2.0`, `MIT`, `CC-BY-4.0`). Otherwise use the
license's own name, linked to its text.
- Pick one of the four values under [Commercial use](#commercial-use). Put any
condition in Notes, in one or two short sentences: the limit, what must be shown, or
where the license does not apply.
- Name bundled parts under a different license, such as a codec, tokenizer or text
encoder, in Notes.
- Set Checked to the date you read the license.
Reference in a new issue View git blame Copy permalink