第三方声明
karanow 建立在开放的模型、库和社区数据之上。以下声明与应用内附带的文件一致。
本站试听曲目
首页试听会播放数段录音，每一段的授权与出处都列在下面。其中几段是美国空军所属乐团的演奏——美国联邦政府作品，属于公共领域，此处署名是出于礼节。其余采用知识共享条款发布；凡授权为 CC BY 的，署名是许可条件而不是礼节。本站的音频是修改过的：karanow 用自己的分离流程把每段录音拆成人声和伴奏两轨。
- Coastline State of Mind — Silver Wings, United States Air Force Band. Public domain, Wikimedia Commons.
- Black Eyes (Trance Extended) — SDM PROJECT. CC BY 3.0, Wikimedia Commons.
- Change My Mind — Max Impact, United States Air Force Band. Public domain, Wikimedia Commons.
- What I'm Sayin' (Streetmix) — MC Shadow. CC0, Wikimedia Commons.
- Skyscrapers — Airmen of Note, United States Air Force Band. Public domain, Wikimedia Commons.
OpenKara Models
The high-quality streaming mode uses the waveform HTDemucs ONNX artifact published by OpenKara Models.
Source: https://github.com/thedavidweng/openkara-models
Pinned artifact: model-v2.1.0/htdemucs.onnx
demucs-mlx
The offline separation mode uses demucs-mlx, an MLX-native port for Apple Silicon.
Copyright (c) 2026 Sam Small
MIT License. Source: https://github.com/ssmall256/demucs-mlx
Demucs
The htdemucs_ft architecture and pretrained model are provided by the Demucs project.
Copyright (c) Meta Platforms, Inc. and affiliates.
MIT License. Source: https://github.com/facebookresearch/demucs
StemgenRT
The optional AI vocal-separation mode uses the HS-TasNet model architecture and model weights distributed by StemgenRT.
Copyright (c) 2026 Axel Delafosse
MIT License
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Source: https://github.com/sweetspotsoundsystem/stemgen-rt
ONNX Runtime
ONNX Runtime is distributed by Microsoft under the MIT License.
Source: https://github.com/microsoft/onnxruntime
SwiftF0
The live singer-pitch and cached reference-pitch paths use the SwiftF0 neural fundamental-frequency model.
Copyright (c) 2025 Lars Nieradzik
MIT License. Source: https://github.com/lars76/swift-f0
Pinned revision: 64700fce8ef39c2970814bf427ac1d75a2f20d72
Pinned model SHA-256: 7e2390db8379cd9e1e2b22828e55b45b57c8559e4c8335678c717dc245c18176
Montserrat
The karanow wordmark and app icon are drawn from Montserrat Black. The glyphs are converted to outlines at build time, so neither the app nor the website ships a font file. The typeface and a copy of this licence live in Design/, which is a source directory: the build takes only Kara.icns from it, so the font is never packaged or served.
Copyright 2011 The Montserrat Project Authors
SIL Open Font License 1.1. Source: https://github.com/JulietaUla/Montserrat