ART vs darktable

ART is a free, open-source RAW image processor for Linux, macOS and Windows. Derived from RawTherapee, it aims for a simpler interface while retaining processing power and quality. Its editing capabilities include layers, batch editing and local adjustments using drawn or parametric masks. It handles metadata and can optionally read and write XMP sidecar files. A plugin system supports custom input and output formats, including true HDR output. For color work, ART supports ACES CLF LUTs through OpenColorIO and CTL scripts through the ACES CTL interpreter. It also includes automatic perspective correction adapted from darktable. User commands can connect external programs; documented AI masking and denoising workflows rely on separate open-source tools, and the project says full AI-tool integration is unlikely in the foreseeable future. ART is licensed under GPL version 3 and comes without a warranty. Stable binaries are linked from GitHub releases, with additional builds listed for Windows, openSUSE, Arch Linux and notarized macOS. Questions and bug reports can be directed to its GitHub issue tracker or discussion board.

darktable is a free photography workflow application for organizing digital negatives and developing raw images. It stores images in a database and provides a zoomable lighttable, while its editing leaves original files unchanged. It imports standard, raw and high dynamic range formats such as JPEG, CR2, NEF, HDR, PFM and RAF. Color tools include automatic display-profile detection on most systems and built-in ICC profiles for sRGB, Adobe RGB, XYZ and linear RGB. Many operations can use OpenCL GPU acceleration, and tethered shooting with live view is supported for some camera brands. Exports include disk storage, email attachments, Piwigo webalbums and HTML-based galleries; many aspects can also be scripted in Lua. Optional AI tools cover object masking, denoising and upscaling, but the Flatpak release does not include them as of September 2026. Downloads are available for Windows, macOS and Linux/Unix options. macOS bundles require version 13.5 or later. The project notes that the learning curve can be steep.