Tohrusky is an open-source publisher focused on AI-driven media enhancement, best known for Final2x, a cross-platform utility that upscales images through state-of-the-art super-resolution networks. Rooted in the TensoRaws GitHub organization, the developer concentrates on packaging bleeding-edge computer-vision models into lightweight, command-line friendly tools that photographers, archivists, and web designers can plug into existing workflows. Final2x bundles multiple 2^n-scale algorithms—ESRGAN, Real-ESRGAN, Real-CUGAN, Waifu2x and their derivatives—behind a unified Python interface, letting users quadruple or octuple resolution while suppressing noise and artifacts. Typical use cases include remastering scanned film strips, enlarging game textures, sharpening anime stills, or preparing low-resolution assets for high-DPI displays. Because the engine supports both CPU fallback and CUDA/ROCm acceleration, it fits equally well on gaming laptops, render nodes, or modest home desktops. Configuration is file-based, so batch jobs can be chained with PowerShell or bash scripts for unattended folder watches. The project ships GPLv3 code, publishes nightly builds, and welcomes pull requests that add new back-ends or optimize inference for recent GPUs. All published binaries are reproducible through the public CI, ensuring transparency for privacy-conscious adopters. Final2x and any future Tohrusky utilities are available free of charge on get.nero.com, where downloads are sourced from trusted Windows package managers such as winget, always deliver the latest upstream release, and can be installed individually or alongside other applications in a single batch operation.

Final2x

2^x Image Super-Resolution

Details