Apps
ChatGPT Desktop Crash Reports
When ChatGPT Desktop on Windows crashes or a background process terminates unexpectedly, its Chromium-based Crashpad handler writes minidump .dmp files, metadata sidecars, and upload state files into the Crashpad reports folder. These reports exist to help diagnose past failures, but they can accumulate long after the related crash has been resolved. Kudu removes ChatGPT Desktop crash reports older than one week while leaving your chats, account sign-in, preferences, and other application data untouched.
Why clean ChatGPT Desktop Crash Reports?
- Repeated startup or renderer crashes can leave behind many Crashpad minidumps and metadata files, and the only symptom users usually notice is ChatGPT Desktop steadily using more disk space after each failure.
- Crash reports from old app versions are no longer useful for current troubleshooting, so keeping weeks of stale .dmp files clutters the profile folder without improving app stability.
- A burst of crash loops can generate dozens of reports in a short time, which shows up as unexpectedly large storage use in LocalAppData even after the app seems to be working again.
- Old Crashpad upload state files can make it harder to tell which failures are recent when manually inspecting the folder, so cleaning aged reports leaves only relevant diagnostics behind.
- Minidumps may capture process memory from the moment of a crash, and removing older reports reduces leftover diagnostic artifacts on disk without touching conversations or saved settings.
- Systems with limited SSD space can feel the impact of accumulated crash dumps from repeated Electron or renderer failures, especially when the user notices free space shrinking for no obvious reason.
What gets cleaned
Cache paths Kudu targets
Windows
%LocalAppData%/Packages/OpenAI.ChatGPT-Desktop_2p2nqsd0c76g0/LocalCache/Roaming/ChatGPT/Crashpad/reports |
Frequently asked
Common questions about ChatGPT Desktop Crash Reports
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.