Standing on the shoulders of open source
Private, on-device photo editing exists because researchers and developers share their work. These are the projects SleekCut builds on. Thank you.
- inpaint-webGPL-3.0lxfater and contributors
Showed that object removal and 4x upscaling can run entirely in the browser. It inspired the SleekCut Object Remover and Image Upscaler. SleekCut uses its own pipeline built on the same open models listed below; no inpaint-web code is included.
- MI-GANMITPicsart AI Research
The inpainting model behind the Object Remover. A fast, mobile friendly model that fills painted areas with matching content.
- Real-ESRGANBSD-3-ClauseXintao Wang and contributors
The general x4v3 super resolution model behind the Image Upscaler and Photo Enhancer.
- MODNetApache-2.0Zhanghan Ke and contributors
Portrait matting model used for the Person mode of the background remover, with fine hair detail.
- U-2-NetApache-2.0Xuebin Qin and contributors
Salient object detection model (U2-Netp) used for the Product or object mode of the background remover.
- Guided Image FilteringResearch paperKaiming He, Jian Sun, Xiaoou Tang
The edge aware filter SleekCut uses to refine cutout edges at full resolution.
- Microsoft
Runs every AI model on your device with WebAssembly and WebGPU.
- Transformers.js model conversionsApache-2.0Joshua Lochner (Xenova) and the Hugging Face community
Browser ready ONNX exports of MODNet. Other ONNX exports come from community uploads on Hugging Face, each used under its original licence.
- heic2anyMITAlex Corvi
Decodes iPhone HEIC photos in the browser.
- fflateMITArjun Barrett
Builds ZIP downloads for batch tools.
- AstroMITThe Astro team
The framework this site is built with.
- Newsreader and Space GroteskSIL Open Font LicenseProduction Type, Florian Karsten
The typefaces used across SleekCut, served through Fontsource.