Codex Temporary Files

Why clean Codex Temporary Files?

• Aborted or crashed Codex runs leave orphaned request payloads, attachment copies, and partial response fragments in .codex/tmp, which shows up as steadily growing disk usage even when no command is running
• Temporary streamed-output files from interrupted sessions can be mistaken for active state on the next launch, causing a command to appear stuck resuming or to pause before starting fresh work
• Large unpacked attachments and prompt expansion scratch files consume home-directory space, leading to obvious "no space left on device" errors during later Codex commands or unrelated developer tools
• Old transient JSON and pipe-spool files can leave behind incomplete artifacts after a failed command, which users notice as half-written output files or leftover temp directories in project folders
• Too many settled temp files in .codex/tmp slow directory scans and cleanup passes, making startup and command teardown feel delayed after heavy CLI use
• On shared or long-lived machines, old temporary command data can preserve copies of prompt inputs and generated intermediate text longer than intended, so cleaning reduces that leftover local footprint without touching saved history or configuration

Common questions about Codex Temporary Files

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.