Pipenv Cache
Pipenv keeps a local cache of downloaded package archives, built wheel files, HTTP responses from Python package indexes, and dependency resolution metadata so installs and lockfile generation do not have to fetch everything again. That cache speeds up repeated pipenv install, sync, and lock operations, but stale artifacts can leave Pipenv reusing outdated package metadata or corrupted downloads. Kudu removes only Pipenv’s cache directories, including package and resolver cache files, without touching virtual environments, Pipfile, Pipfile.lock, project code, accounts, or credentials.
Why clean Pipenv Cache?
- A partially downloaded sdist or wheel in Pipenv’s cache can trigger hash mismatches or repeated install failures when pipenv install keeps reusing the same bad file
- Cached package index responses can leave Pipenv seeing outdated dependency metadata, so pipenv lock reports version conflicts even though newer compatible releases exist upstream
- Old resolver cache entries can make lock generation behave inconsistently after package releases change, showing long dependency resolution times or a lockfile that will not refresh cleanly
- Reused cached artifacts from a previous build environment can cause install errors during wheel unpacking or metadata preparation, which shows up as failed dependency installs on one machine but not another
- The cache can grow with multiple copies of wheels, source tarballs, and HTTP responses, leading to unexpectedly high disk usage under the user cache folder while projects themselves stay small
- After clearing stale cache files, Pipenv simply re-downloads needed packages and metadata, which often fixes repeated install or sync errors without affecting virtual environments, project files, or saved credentials
Cache paths Kudu targets
Windows
%LocalAppData%/pipenv/pipenv/Cache |
macOS
~/Library/Caches/pipenv |
Linux
~/.cache/pipenv |
Common questions about Pipenv Cache
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.