Skip to content

Make videos on the computer you already own. FreeVideo runs MiniMax H3 in as little as 8 GB of VRAM and 16 GB of RAM, and adapts its acceleration path to your hardware.

8.3 KB
Raw Blame

Third-party notices

FreeVideo code uses Apache-2.0. Its dependencies and models retain their own licenses.

Component License
Three.js r160, © 2010–2023 three.js authors (bundled whitebox renderer) MIT
Depth Anything V2 Small, © 2024 Depth Anything authors; DINOv2 © Meta Platforms, Inc. and affiliates. Small-only DPT/DINOv2 inference and preprocessing are adapted and bundled; Small weights are an optional download pinned at 03876f8651c73a60fe4c2c48294e09fcb6838fcf. Apache-2.0 (code and Small weights)
Qt, PySide6 and Shiboken6 LGPL-3.0; component notices
PyInstaller bootloader GPL-2.0 with the bootloader exception
CPython / psutil PSF / BSD-3-Clause
QuickJS Python binding 1.19.4 (author: Petter Strandmark) and QuickJS (© 2016–2021 Fabrice Bellard, © 2017–2021 Charlie Gordon) MIT notices
ComfyUI, including the H3 text encoder code GPL-3.0
VDN-H3, Diffusers, SageAttention Apache-2.0
VDN-H3 and H3 model weights MiniMax H3 Community License
MiniMax H3 latent upscaler MIT
FreeToken prompt VLM operators, Qwen3-VL-4B-Instruct optional weights Apache-2.0
Qwen-Image-2.1 Turbo image generation models, including Qwen3-VL-8B and VAE (optional download; notice installed with the models), © 2026 Hangzhou Tongyi Laboratory Technology Co., Ltd. All Rights Reserved. Qwen RESEARCH LICENSE AGREEMENT
sherpa-onnx 1.13.8 speech runtime, built by FreeVideo without text-to-speech, ALSA and PortAudio (modified build; __init__.py edited; source snapshot 11afbd00 kept by FreeVideo; optional download for automatic subtitles, build note). Includes statically linked kaldi-native-fbank, kaldi-decoder, kaldifst, OpenFst, simple-sentencepiece (Apache-2.0), KissFFT (BSD-3-Clause), pybind11 (BSD-3-Clause), nlohmann/json (MIT), hclust-cpp and Darts-clone (BSD-2-Clause), ThreadPool (Zlib). The Windows build also contains msvcp140.dll and msvcp140_1.dll (Microsoft Visual C++ 2022 runtime 14.44.35211, © Microsoft Corporation), unmodified and redistributed as permitted by Microsoft for its redistributable files. Apache-2.0
CosyVoice code adapted for dubbing and Fun-CosyVoice3-0.5B-2512 weights, © 2024 Alibaba Inc (optional download); includes MIT-licensed parts of Matcha-TTS, HiFi-GAN, x_transformers and Snake Apache-2.0; MIT notices
ONNX Runtime 1.28.2 (CPU), © Microsoft Corporation; its third-party notices are shipped inside the sherpa-onnx wheel MIT
Eigen 5.0.1, compiled into sherpa-onnx and ONNX Runtime; source: https://gitlab.com/libeigen/eigen/-/tree/5.0.1 MPL-2.0
Silero VAD v5.0 voice-activity model, © 2020-present Silero Team (optional download with subtitles, license file downloaded with the model; the onnx-community export is downloaded with the music model) MIT
ACE-Step 1.5 XL turbo music model, © 2026 ACEStep (optional download for music; inference subset adapted in freevideo_engine/music/acestep_model.py) MIT
Bandit Plus cocktail-party separation model, weights by Watcharasupat et al. (optional download for sound separation, attribution required, non-commercial); code from Bandit, © the Bandit authors, reduced to an inference subset in freevideo_engine/stems/bandit/ (modified) CC BY-NC 4.0 (weights), Apache-2.0 (code)
vector-quantize-pytorch, © 2020 Phil Wang (residual FSQ decoding adapted in freevideo_engine/music/acestep_model.py) MIT
Qwen3-Embedding-0.6B, © Alibaba Cloud (text encoder of the music model, downloaded with it) Apache-2.0
NVIDIA NeMo stt_en_conformer_ctc_small, © NVIDIA, exported to ONNX and quantized to int8 by the sherpa-onnx project, changed from the original; not endorsed by NVIDIA (optional download for English subtitles; notice) CC-BY-4.0
OpenAI Whisper tiny, © 2022 OpenAI, exported to ONNX and quantized to int8 by the sherpa-onnx project (optional download for spoken-language detection) MIT
SenseVoiceSmall by FunASR / FunAudioLLM (Alibaba Group), converted to ONNX and quantized to int8 by the sherpa-onnx project, used unchanged under its original name (optional download for multilingual subtitles; notice) FunASR Model Open Source License Agreement v1.1
Lucide icons (image editor: eyedropper, hand, crop) ISC
Noto Sans CJK SC and Noto Sans Mono CJK SC 2.004, © 2014-2021 Adobe: subsets renamed FreeVideo Sans SC and FreeVideo Mono SC (image editor text, editor text and subtitles / 剪辑视图的文字和字幕, share card, Linux launcher) SIL OFL 1.1
Noto Serif CJK SC 2.003, © 2017-2024 Adobe: subset renamed FreeVideo Serif SC (image editor text, editor text and subtitles / 剪辑视图的文字和字幕) SIL OFL 1.1
LXGW WenKai v1.522, © 2021-2026 LXGW, © 2020 The Klee Project Authors: subset renamed FreeVideo Kai SC (image editor text, editor text and subtitles / 剪辑视图的文字和字幕) SIL OFL 1.1

Original launcher license texts are included in Windows packages. Other runtime dependencies are listed in constraints, with their license information in the installed packages.

The fonts in web/fonts/ are built by scripts/build_web_fonts.py from the pinned sources above; each file keeps its original copyright notice and licence in its name table. FreeVideoBox.woff2 is FreeVideo's own.

decode_stream.py, lora_cache.py and reference_sampler.py include adaptations of upstream code; their source headers retain the attribution.

prompt_vlm/core.py adapts FreeToken's merged MLP/QKV and bounded vision operators from revision 555efd89447232a555e05d187256b76a51c1aaa0. Its header describes the changes and integration boundary.