GIMP-ML brings machine learning directly into GIMP, the free and open-source image editor. It is a set of plug-ins that add AI-powered tools – including image colorization, deblurring, dehazing, super-resolution and semantic segmentation – to GIMP’s menus, so photo editors can apply modern deep learning models without leaving their normal editing workflow. The project has around 1,600 stars on GitHub.
Most AI photo tools live in separate web apps, which breaks an editor’s flow and forces images through third-party services. GIMP-ML instead installs a Python plug-in layer inside GIMP: select a layer, choose an AI operation, and the model runs locally and returns the processed result as a new layer. The collection covers common restoration and enhancement tasks, from turning black-and-white photos into color, to removing haze and blur, enlarging images and generating depth maps for effects. It demonstrated early how open editors could absorb AI features, years before AI menus became standard.
GIMP-ML is free and open source under the MIT license, and GIMP itself is free as well.
Photographers and GIMP users who want practical AI restoration and enhancement tools inside their existing open-source editor.
GIMP-ML’s plug-ins are built on PyTorch models that the author trained or adapted for each task, and the installation bundles them as a single GIMP plug-in pack for Linux, Windows and macOS. The project showed that an open-source editor could gain AI features comparable to commercial software through community plug-ins, and its menu-based approach – pick a layer, run a model, get a layer back – remains a clean pattern for editor AI.