Licences

Open-source software included in FrameAnvil

FrameAnvil is built with, and includes, the following open-source software. We’re grateful to everyone who works on it.

ComponentUsed forLicence
FFmpeg and FFprobeReading media, previews, waveforms and every exportGNU GPL version 3 or later
Electron (with Chromium and Node.js)The application itselfMIT (Chromium: BSD-style and others)
PDF.jsImporting Illustrator and PDF artworkApache License 2.0
PDFiumReading the layers of Illustrator and PDF artwork, and drawing CMYK artworkBSD 3-Clause

FFmpeg

FrameAnvil includes unmodified builds of the ffmpeg and ffprobe programs from the FFmpeg project, licensed under the GNU General Public License, version 3 or later. They are separate programs that FrameAnvil runs, and their licence text is included with every copy of FrameAnvil.

You may get the complete corresponding source code of these programs from ffmpeg.org. On request, we will also provide it ourselves, for at least three years from the date you received FrameAnvil, for no more than the cost of providing it: write to legal@frameanvil.com and tell us the FrameAnvil version you have.

FFmpeg is a trademark of Fabrice Bellard, originator of the FFmpeg project.

Electron, Chromium and Node.js

Electron is released under the MIT licence. It contains Chromium and Node.js, each made of many components under their own licences; the full list is included with FrameAnvil in the file LICENSES.chromium.html.

PDF.js

PDF.js is Copyright Mozilla Foundation and released under the Apache License, Version 2.0.

PDFium

PDFium is Copyright 2014 The PDFium Authors and released under the BSD 3-Clause License. FrameAnvil uses its WebAssembly build from @hyzyla/pdfium (MIT licence), which is built with FreeType (FreeType License), Little CMS (MIT), OpenJPEG (BSD 2-Clause), libjpeg-turbo (IJG and BSD 3-Clause), libpng, zlib, Anti-Grain Geometry 2.3 and Abseil (Apache License 2.0), each under its own licence.

Transformers.js and ONNX Runtime

Automatic captions, background removal and upscaling run on Transformers.js (Copyright Hugging Face, Apache License, Version 2.0), ONNX Runtime Web and ONNX Runtime for Node.js (Copyright Microsoft Corporation, MIT licence). On Windows, ONNX Runtime includes DirectML and the DirectX Shader Compiler (Copyright Microsoft Corporation), redistributed under their licences.

Whisper

The speech-recognition model used for captions is OpenAI's Whisper (Copyright 2022 OpenAI, MIT licence), in the version converted for Transformers.js by Xenova. It's included with FrameAnvil.

MODNet

The model used for background removal in videos (and in photos, if you choose) is MODNet (Copyright Zhanghan Ke and contributors, Apache License, Version 2.0), in the version converted for Transformers.js by Xenova. It's included with FrameAnvil.

Image-Adaptive 3D LUT

Color › Auto uses the tone model from “Learning Image-adaptive 3D Lookup Tables for High Performance Photo Enhancement in Real-time” (Hui Zeng, Jianrui Cai, Lida Li, Zisheng Cao and Lei Zhang), released under the Apache License, Version 2.0. It's included with FrameAnvil.

BiRefNet

The model used for background removal in photos is BiRefNet lite (Copyright 2024 ZhengPeng, MIT licence), in the version converted for Transformers.js by onnx-community. It's included with FrameAnvil.

Real-ESRGAN

Upscale uses the Real-ESRGAN models realesr-general-x4v3 and realesr-animevideov3 (Copyright 2021 Xintao Wang, BSD 3-Clause licence), converted to ONNX for FrameAnvil. They're included with FrameAnvil.

GFPGAN

Restore faces uses GFPGAN v1.4 (Copyright 2021 Tencent ARC Lab, Apache License, Version 2.0), in the ONNX version published by the FaceFusion project. It's included with FrameAnvil.

YuNet

Faces are found for that with YuNet (Copyright Shiqi Yu and contributors, from the OpenCV Zoo, MIT licence). It's included with FrameAnvil.

Questions

Write to legal@frameanvil.com.