Skip to main content

aiclean

Scan and safely clean local caches left behind by AI coding tools — Claude Code, Cursor, VSCode, Ollama, npm/pnpm/yarn, uv/pip, HuggingFace, PyTorch, Playwright, Hermes, and more.

AI dev tools are notorious for quietly eating disk space: Ollama models, Claude Code project logs, HuggingFace model caches, browser binaries for test automation... it adds up to tens of gigabytes without anyone noticing. aiclean finds it, shows you exactly where it is, and only deletes what you explicitly confirm.

Install

pip install aiclean
# or run without installing:
pipx run aiclean

Usage

# Scan only (read-only, safe to run anytime)
aiclean

# See all known cache locations this tool checks
aiclean --list

# Preview what would be deleted (no changes made)
aiclean --clean --dry-run

# Interactively confirm each cache before deleting
aiclean --clean

# Clean specific targets without prompts (careful!)
aiclean --clean --only ollama-models npm-cache --yes

Example output

      SIZE  NAME                     PATH
----------------------------------------------------------------------
    12.4GB  ollama-models            /Users/you/.ollama/models
     3.1GB  claude-code-projects     /Users/you/.claude/projects
     1.8GB  huggingface-cache        /Users/you/.cache/huggingface
   420.0MB  npm-cache                /Users/you/.npm
   180.2MB  uv-cache                 /Users/you/.cache/uv
----------------------------------------------------------------------
    17.9GB  TOTAL (5 caches found)

Run with --clean to interactively remove caches (add --dry-run to preview safely).

Safety

  • Scan mode never touches the filesystem — it only computes sizes.
  • --clean without --yes asks per-target before deleting anything.
  • --dry-run combined with --clean shows exactly what would be removed, with zero deletions.
  • Targets are well-known, documented cache directories only — aiclean never does a generic recursive delete of arbitrary paths.

What it scans for

Run aiclean --list for the full, current list. Broadly: editor/IDE caches (Cursor, VSCode), AI coding agent caches (Claude Code, Hermes), local model caches (Ollama, HuggingFace, PyTorch Hub), and package manager caches (npm, pnpm, yarn, uv, pip) — since these are the biggest and most overlooked space users of an AI-assisted dev setup.

Contributing

Issues and PRs welcome — especially "please add support for X cache location" requests, since that's exactly the kind of real-world signal this project is looking for.

License

MIT

Metadata

Release files for aiclean 0.1.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for aiclean 0.1.0
File Size Uploaded Provenance Actions
aiclean-0.1.0.tar.gz 4.7 kB Without provenance Details for aiclean-0.1.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for aiclean 0.1.0
File Interpreter ABI Platform Actions
aiclean-0.1.0-py3-none-any.whl Python 3 none any Details for aiclean-0.1.0-py3-none-any.whl

Total release size: 10.0 kB

Release files / aiclean-0.1.0.tar.gz

Download URL aiclean-0.1.0.tar.gz
Size 4.7 kB
Tags Source
SHA-256 checksum
How to use checksums
32d04edea0df50b2588117f1a3fb2ad563d0ba228339fe7e2739cf6c88cc58bc
BLAKE2b-256 checksum
How to use checksums
e9559a1f2219b467a21bb6302eade287a00576a318615c4fced9e7a14b9a536e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.9.6

Release files / aiclean-0.1.0-py3-none-any.whl

Download URL aiclean-0.1.0-py3-none-any.whl
Size 5.2 kB
Tags Python 3
SHA-256 checksum
How to use checksums
a381721d816455bbacb863755e140724d6d4135b922b5b1358b085905a2d868a
BLAKE2b-256 checksum
How to use checksums
d1aaca7cfd5a7219595bd8aadbe0afcbfe8cfebe6d98817adf2498a451950540
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.9.6

Security for aiclean 0.1.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

0.1.0 This release

2 release files No provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page