← All Tools

Open-Source Credits

Our free tools run entirely in your browser, and they stand on excellent open-source work. These are the libraries and AI models each tool ships, their authors, and their licenses. Thank you to every maintainer below.

AI runtimes & models

ONNX Runtime Web

© Microsoft Corporation · MIT License · github.com/microsoft/onnxruntime
The in-browser AI engine behind the Background Remover.

BEN2 (Background Erase Network)

© PramaLLC · MIT License · huggingface.co/PramaLLC/BEN2 (ONNX export: onnx-community)
The high-quality segmentation model in the Background Remover and Passport Photo Maker.

MODNet

© Zhanghan Ke and MODNet contributors · Apache License 2.0 · github.com/ZHKKKe/MODNet (ONNX export: Xenova)
The fast portrait-matting model in the Background Remover and Passport Photo Maker.

4x-ClearRealityV1

© Kim2091 · Apache License 2.0 · huggingface.co/Kim2091/ClearRealityV1
The super-resolution models (Detail and Soft variants, built on the SPAN architecture, Apache-2.0) behind the AI Photo Enhancer and Image Upscaler.

mp4box.js

© Telecom ParisTech/TSI/MM/GPAC, Cyril Concolato · BSD 3-Clause License · github.com/gpac/mp4box.js
Reads MP4 and MOV containers in the Video Compressor.

mp4-muxer

© Vanilagy · MIT License · github.com/Vanilagy/mp4-muxer
Writes the fast-start MP4 the Video Compressor produces.

NAFNet

© megvii-model · MIT License · github.com/megvii-research/NAFNet (ONNX export: OpenCV Zoo, MIT; weights re-quantised per channel by AccioPix)
The motion-deblur model behind Unblur Image.

NIMA (idealo image-quality-assessment)

© idealo internet GmbH · Apache License 2.0 · github.com/idealo/image-quality-assessment
The aesthetic-scoring model behind the Top Photo Picker's appeal ranking (converted to ONNX by AccioPix).

jSquash (Squoosh codecs for the web)

© jamsinclair and the Squoosh contributors · Apache License 2.0 · github.com/jamsinclair/jSquash
The WebAssembly builds of the encoders behind the Photo Compressor: MozJPEG (© Mozilla, IJG and libjpeg-turbo licences), libwebp (© Google, BSD-3), libavif with the AOM encoder (BSD-2 and the AOM licence), oxipng (MIT) and the image-png decoder (MIT/Apache-2.0).

MediaPipe Tasks Vision

© Google LLC · Apache License 2.0 (runtime and models) · ai.google.dev/edge/mediapipe
The face detector (BlazeFace) and Face Landmarker models behind the Closed-Eye Checker and the Passport Photo Maker's auto-framing.

Libraries

heic2any

© Alex Corvi · MIT License · github.com/alexcorvi/heic2any
HEIC decoding in the HEIC to JPG, HEIC to PNG and HEIC to PDF converters. Bundles the libheif and libde265 codecs (© struktur AG / Dirk Farin, LGPL-3.0-or-later) and gifshot (© Yahoo Inc., MIT).

exifr

© Mike Kovařík, Mutiny.cz · MIT License · github.com/MikeKovarik/exifr
Metadata parsing in the EXIF Viewer & Remover and the RAW converters.

jsPDF

© James Hall, yWorks GmbH · MIT License · github.com/parallax/jsPDF
PDF generation in the HEIC to PDF converter.

JSZip

© Stuart Knightley, David Duponchel, Franz Buchinger, António Afonso · MIT License (dual-licensed MIT / GPL-3.0-or-later; used here under MIT) · github.com/Stuk/jszip
ZIP downloads in every tool. Includes pako (© Vitaly Puzrin and Andrei Tuputcyn, MIT).

browser-image-compression

© Donald Chan · MIT License · github.com/Donaldcwl/browser-image-compression
Image compression in the album-export flow on the tool pages.

The copies of these projects served from this site retain their original license notices where the published builds include them; full license texts are available at the links above. Everything else in the tools — the culling logic, RAW preview extraction, blink scoring and the pages themselves — is AccioPix's own work.