Apps
T3 Code Partitions
T3 Code stores per-partition preview caches under its Partitions directory so each isolated workspace can render web-based panels and embedded tools without rebuilding everything on every launch. Those folders commonly accumulate Chromium-style HTTP cache entries, GPUCache shader data, Code Cache V8 bytecode, blob storage, and temporary service worker files created by partitioned preview sessions. Kudu removes only these disposable preview-partition caches while leaving partition sessions, site storage, workspace state, accounts, settings, and passwords untouched.
Why clean T3 Code Partitions?
- Stale HTTP cache files inside a partition can keep serving outdated preview assets, so an embedded app or local preview loads the wrong JavaScript or CSS and you see old UI behavior after a rebuild
- Invalidated Code Cache V8 bytecode after a T3 Code or Chromium runtime update can make the first open of a preview panel unusually slow until the bytecode cache is rebuilt
- Old GPUCache shader data from a previous graphics driver version can cause black preview panes, flickering canvases, or stutter when a WebGL-based tool opens inside a partition
- Corrupted blob storage or temporary service worker cache data can leave an embedded login or preview stuck on a spinner even though the underlying project and partition data are still present
- Oversized per-partition caches accumulate across many workspaces, so roaming AppData grows unexpectedly and T3 Code takes longer to open preview-backed panes while Windows antivirus scans thousands of small cache files
- Broken cached responses in one partition can affect only that isolated preview environment, which shows up as one workspace failing to load assets while other partitions in T3 Code still work normally
What gets cleaned
Cache paths Kudu targets
Windows
%AppData%/t3code/Partitions |
Frequently asked
Common questions about T3 Code Partitions
Related cleaners
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.