T3 Code Logs
T3 Code writes rotating diagnostic logs under its userdata logs directory to record extension host startup, language server handshakes, terminal integration events, renderer warnings, crash traces, and update diagnostics. These plain-text session logs accumulate as the app launches, reloads windows, installs extensions, and reconnects background services, but older files stop being useful once the relevant session has passed. Kudu removes only T3 Code log files older than one week from ~/.t3/userdata/logs, without scanning or touching attachments, secrets, runtime state, settings, projects, or account data.
Why clean T3 Code Logs?
- Weekly log buildup from repeated launches, extension reloads, and language server restarts quietly consumes disk space, and users usually notice only when the T3 userdata folder becomes unexpectedly large
- Crash loops and failing extensions can generate dense renderer and extension-host traces, leaving behind hundreds of megabytes of old logs even after the actual problem is fixed
- Verbose terminal and update diagnostics from past sessions make the logs folder harder to inspect manually, so recent errors get buried under stale files when troubleshooting a current issue
- Large collections of old text logs increase the amount of filesystem clutter antivirus, backup, and file indexing tools must scan, which users may notice as slower backups or longer searches in the userdata directory
- Because these are historical diagnostics rather than live caches, keeping months of expired session logs offers little benefit while making disk usage grow after every window reload and extension install
- Old log sets often preserve repeated copies of the same startup warnings and reconnect traces, so cleaning them reduces noise without affecting current workspaces, saved files, settings, or credentials
Cache paths Kudu targets
Windows
~/.t3/userdata/logs |
Common questions about T3 Code Logs
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.