Apps

Upscayl

Upscayl stores several Electron runtime caches under its Windows profile, including Chromium HTTP cache files in Cache_Data, V8 compiled script data in Code Cache, GPUCache entries, and DawnCache shader and pipeline blobs used by its GPU rendering path. These files speed up the app shell, previews, and model-selection interface after the first run, while the logs folder accumulates settled diagnostic logs from launches and upscaling sessions. Kudu removes those temporary caches and logs without touching your images, saved output, app settings, or any preserved image databases.

Why clean Upscayl?

  • Invalidated V8 compiled code after an Upscayl update can make the window open slowly or feel unresponsive on the first few launches until Code Cache is rebuilt
  • Stale GPUCache data left over after a graphics driver update can cause black panels, missing preview elements, or flickering in the Electron interface until the GPU cache is cleared
  • Outdated DawnCache shader and pipeline blobs can conflict with a newer WebGPU or Direct3D stack, showing up as stutter, delayed preview drawing, or a window that appears blank before rendering catches up
  • Corrupted Chromium Cache_Data entries can make bundled web assets load incorrectly, leading to broken styling, missing icons, or buttons in the Upscayl UI that do not paint properly
  • Settled logs from repeated launches and failed upscale attempts keep growing in the background, so disk usage increases even though your actual input and output images have not changed
  • A damaged cache can trap Upscayl in repeated slow starts where the interface hangs on launch, then behaves normally only after temporary runtime files are regenerated
What gets cleaned

Cache paths Kudu targets

Windows

%AppData%/Upscayl/Cache/Cache_Data
%AppData%/Upscayl/Code Cache
%AppData%/Upscayl/GPUCache
%AppData%/Upscayl/DawnCache
%AppData%/Upscayl/logs
Frequently asked

Common questions about Upscayl

Free & open source

Download Kudu and reclaim your disk space.

Available on Windows, macOS, and Linux. No account required, no feature gates, no telemetry without consent. All cleaning targets are open source and community-auditable.