Open source attributions

ToolsAllInOne is built on other people's work. Here is what it uses and under which licence.

All processing happens on your device using the projects below. Each remains the property of its authors under its own licence. The full copyright and licence notices for every component served from this site are in THIRD-PARTY-NOTICES.txt, and those for the native libraries compiled into the WebAssembly engines, plus the ONNX Runtime and Jinja code bundled inside Transformers.js, are in EMBEDDED-NOTICES.txt. Where a component is marked Apache 2.0, that link is the full licence text, and it covers every Apache-licensed component on this page, including those loaded from public CDNs.

FFmpeg

Audio and video conversion uses FFmpeg n5.1.4 compiled to WebAssembly as @ffmpeg/core 0.12.6, built by the ffmpeg.wasm project from the build scripts at its v0.12.10 tag. Your browser downloads that binary from unpkg.com or cdn.jsdelivr.net; it is not served from this site. It is configured with --enable-gpl and links x264, x265 and the libraries below, which makes the combined binary licensed under the GNU General Public License, version 2 or later.

  • FFmpeg — ffmpeg.org, GPL-2.0-or-later in this configuration.
  • x264 — videolan.org, GPL-2.0-or-later. x265 — videolan.org, GPL-2.0-or-later.
  • LAME (MP3) — LGPL-2.0-or-later. libvpx (VP8/VP9) — BSD-3-Clause. Opus — BSD-3-Clause. libogg, libvorbis and libtheora — BSD-3-Clause. libwebp — BSD-3-Clause. zlib — zlib licence.
  • FreeType — FreeType License or GPL-2.0-or-later. FriBidi — LGPL-2.1-or-later. HarfBuzz — MIT. libass — ISC. zimg — WTFPL.
  • The ffmpeg.wasm loader (@ffmpeg/ffmpeg 0.12.10 and @ffmpeg/util 0.12.0, served from this site) — github.com/ffmpegwasm/ffmpeg.wasm, MIT licence, copyright (c) 2019 Jerome Wu.
  • Build source for this exact build: our copy of the ffmpeg.wasm v0.12.10 repository — its Dockerfile, build scripts and modified FFmpeg command-line sources. That build fetches FFmpeg n5.1.4 and the libraries it links (x264, x265, libvpx, LAME, Ogg, Theora, Opus, Vorbis, zlib, libwebp, FreeType, FriBidi, HarfBuzz, libass and zimg) from their own repositories at the versions pinned in those scripts. Upstream: github.com/ffmpegwasm/ffmpeg.wasm/tree/v0.12.10.

Codecs and patents

Software licences and patents are separate things, and the GPL covers only the former. Some video and audio formats are covered by patents held by third parties, independently of the software that implements them:

  • H.264 (used for MP4 output) and HEVC/H.265 are patent-encumbered formats. Encoding happens on your own device rather than on our servers. This site never encodes HEVC: x265 is part of the upstream FFmpeg build, but no tool calls it. If you open an HEVC video (the default on recent iPhones), FFmpeg's built-in HEVC decoder reads it on your device, and the few tools that copy the video stream without re-encoding it (such as remove audio, and add audio to video) leave it as HEVC in the output.
  • AAC audio is likewise covered by patents held by third parties.
  • MP3 patents expired in 2017. WebM with VP8 video and Vorbis audio, FLAC, PNG and WebP are free of known patent licensing requirements — if that matters for your use, prefer those formats. Opus and VP9, which browsers use for screen recordings, are subject to third-party patent pool claims.

PDF

Data, text and code

OCR and on-device AI

These run on your device. The engines marked "from a CDN" are fetched from a public CDN the first time you use the tool; the models are fetched from Hugging Face.

  • Tesseract.js 5.1.1 — github.com/naptha/tesseract.js, Apache 2.0, wrapping the Tesseract OCR engine (Apache 2.0), with its English recognition data (tesseract.js-core 5.1.1 and @tesseract.js-data/eng 1.0.0). Served from this site. The compiled engine embeds Leptonica, libpng, libjpeg-turbo, libtiff, giflib, libwebp and zlib; this software is based in part on the work of the Independent JPEG Group. Used by the image-to-text and PDF OCR tools.
  • ONNX Runtime Web 1.17.0 — Microsoft, MIT licence. From a CDN. Runs the background remover's model.
  • U²-Net — github.com/xuebinqin/U-2-Net, Apache 2.0. The salient-object model behind the background remover, served from this site.
  • MediaPipe Tasks Vision 0.10.14 — Google, Apache 2.0. Served from this site, with the BlazeFace short-range face-detection model, © Google LLC, Apache 2.0. Used by the automatic face blur.
  • Transformers.js 2.17.2 — github.com/xenova/transformers.js, Apache 2.0. Served from this site; its WebAssembly runtime comes from a CDN. Runs the models below, which are ONNX conversions published by Xenova; where a conversion carries no licence tag of its own, the upstream model's licence applies.
  • Whisper (tiny.en) — OpenAI, Apache 2.0. Used for transcription and automatic subtitles.
  • Swin2SR — caidas/swin2SR-classical-sr-x2-64, Apache 2.0. Used for AI upscaling.
  • SpeechT5 and its HiFi-GAN vocoder — Microsoft, MIT licence. The voice is a single speaker embedding (CMU ARCTIC speaker "slt", from Matthijs/cmu-arctic-xvectors, MIT licence, derived from the CMU ARCTIC recordings), served from this site. Used by text to speech.
  • Opus-MT translation models by Helsinki-NLP (Language Technology Research Group, University of Helsinki) — huggingface.co/Helsinki-NLP. Fifteen language pairs are under Apache 2.0. The en→de, ru→en and zh→en models are under CC BY 4.0 and are provided without warranty. Used by the translator. Models under non-commercial licences are deliberately not used anywhere on this site.

Trademarks

QR Code is a registered trademark of DENSO WAVE INCORPORATED. The name is used on this site only to describe what the generator produces — there is no affiliation with, sponsorship by or endorsement from DENSO WAVE. Other product, company and format names are the trademarks of their respective owners; see section 8 of the terms.

Typefaces

Icons

The interface icons are taken from or based on Feather by Cole Bemis — feathericons.com, MIT licence, copyright (c) 2013-2023 Cole Bemis. The full notice is in THIRD-PARTY-NOTICES.txt.

Images

Converting between PNG, JPG, WebP, GIF, BMP and ICO uses your browser's built-in image APIs plus this site's own small BMP and ICO writers, with no third-party code. Saving images as PDF uses pdf-lib, and opening TIFF files uses FFmpeg, both listed above.

If you believe something here is attributed incorrectly, please write to [email protected].