Package cache is different from an environment
pip and uv can keep downloaded wheels and build artifacts. Conda distributions keep package caches as well as installed environments. Deleting the wrong tree can remove the interpreter and installed dependencies you are trying to keep.
Ask each tool for its actual location
python -m pip --version
python -m pip cache dir
python -m pip cache info
uv --version
uv cache dir
conda --version
conda info
conda env listRun pip through the Python environment you use. Custom settings can move these caches; do not rely on one hard-coded Library path for all installations. Inspect the package-cache paths reported by Conda before cleanup.
Clean pip downloads
python -m pip cache purgeThis clears pip’s download/build cache, not the packages already installed in your environment. Future installs may download and build again. If you depend on an offline wheel cache or an unavailable private registry, keep the required artifacts first.
Use uv’s own cache commands
uv cache prune
# To intentionally clear all cache entries instead:
uv cache cleanUse these commands rather than removing uv cache files by hand. Prune and clean have different scope; consult the installed help before choosing. They do not mean deleting arbitrary project .venv folders or all uv-managed Python toolchains. A project environment is a separate cleanup decision.
Preview Conda cleanup
conda clean --all --dry-run
# After reviewing the proposed removals:
conda clean --allReview the dry run and confirmation. Do not substitute --force-pkgs-dirs: it can break environments using links to the package cache. Leave envs directories, named environments, configuration and credentials alone. Mamba/Micromamba distributions can use different root prefixes; use their own info/help before cleanup.
Match the action to your goal
| Goal | Action | Keep |
|---|---|---|
| Faster repeated installs | Retain frequently used downloads | Private/offline packages |
| Free download-cache space | pip/uv/Conda cleanup commands | Installed environments and project lockfiles |
| Retire an old environment | Review that environment separately | Active interpreters and custom data |
DevCleaner separates Conda Package Cache (Warning, opt-in) from standalone solver caches (Safe). Its Python and keep-newest scanners have separate scopes. Project Artifacts can surface stale .venv directories, which need a deliberate reinstall decision. See the scanner reference and version cleanup policies.
FAQ
- Does pip cache purge uninstall packages?
- It clears the pip cache, not packages already installed in an environment. Later installs may need downloads or rebuilds.
- Can I delete a Conda pkgs directory by hand?
- Prefer conda clean with a dry run. Broad package-directory removal can break linked environments; do not confuse pkgs with envs.
References and scope
Reviewed against DevCleaner 1.14.5 and the references below on October 4, 2026. Locations can change with tool versions and custom settings. The examples are inspection steps and deliberate cleanup actions, not a measured claim about how much space your Mac will recover.
Review the space before cleaning
DevCleaner separates built-in Safe caches from Warning and Danger items. Safe Quick Clean is free; deep cleanup requires Pro on new Free installations. See Free and Pro features.
Download DevCleaner ↓