conda-clean

Delete all temporary or unused files verbosely and say yes to all confirmations

$ conda clean [[-avy|--all --verbose --yes]]

$ conda clean [[-itl|--index-cache --tarballs --logfiles]]

$ conda clean [[-c|--tempfiles]] [path/to/tempfiles]

$ conda clean [[-p|--packages]]

$ conda clean [[-f|--force-pkgs-dirs]]

$ conda clean [[-h|--help]]

conda clean [OPTIONS]

--all, -a
    Remove index cache, lock files, unused cache packages, and tarballs (equivalent to -p -t -s -i)

--packages, -p
    Remove unused packages from writable cache directories

--source-cache, -s
    Remove extraction caches for unlinked packages

--tarballs, -t
    Remove cached tar files

--index-cache, -i
    Remove cached index files from all index caches

--json
    Return JSON report listing what would be removed

--dry-run
    Report actions without changes

--force-pkgs-dirs
    Use hard links for inter-pkgs-dir linking (symlinks otherwise)

--force-preferred
    Use hard links for packages in preferred index

--log-level <LEVEL>
    Set logging verbosity (debug, info, warnings, error, critical; default: info)

-h, --help
    Show help message and exit

-V, --version
    Show conda-clean version and exit

The conda clean command frees disk space by removing unused packages, tarballs, index caches, source caches, and lock files accumulated by Conda.

Conda, a cross-platform package and environment manager popular for Python and scientific computing, downloads packages into cache directories (<pkgs_dir>/cache). These caches persist even after installation or environment removal, leading to substantial disk usage growth over time—often gigabytes.

conda clean provides targeted cleanup options: remove only unused packages (--packages), tarballs (--tarballs), extraction caches (--source-cache), or index files (--index-cache). Use --all for comprehensive cleanup. Dry-run (--dry-run) and JSON (--json) modes allow previewing actions without changes, aiding safe usage.

It's essential for users with limited storage, multiple environments, or heavy package activity. Post-cleanup, Conda redownloads only as needed, preserving functionality while optimizing space. Always run in base environment unless specified.

Cleanup is irreversible; Conda redownloads removed files on demand. Use --dry-run first. Run in base environment to affect all pkgs_dirs. May require admin privileges for some caches.

conda clean --all (full cleanup)
conda clean -p (packages only)
conda clean --dry-run --json (preview report)

Typically ~/.conda/pkgs or $CONDA_PKGS_DIR. Check with conda info.

Introduced in Conda 4.4 (2017) to address cache bloat complaints. Enhanced in later versions (e.g., 4.7+) with --json, --source-cache, and hard-link options for efficiency.

conda(1)