Apps to vibe code this weekend instead of a disk cleaner

Writing a good disk cleaner is a much bigger job than it looks. Deleting a file takes a second. Knowing which ones are safe to delete, across every app, cache, and edge case, while earning enough trust that someone lets you touch their drive, is the hard part. Pulling a plant out of the ground takes a second too. Knowing which ones are weeds is the gardening.

There is a second reason to walk away. The world already has a thousand cleaners, and a new one is an applicant for a job that stopped taking applications. Nobody has taken the listing down.

Respect the scope

Realizing a project is too big is not quitting. It is the same skill that keeps good software small: reading the scope honestly and believing what it says. A cleaner you correctly decide not to build has already protected more files than most cleaners ever will.

Here is what to build instead. Each is small enough to finish, which is the entire point.

The AI classics

A thin wrapper around an LLM is a weekend project, which is why everyone ships these. The wrapper is thin. The confidence is not.

The productivity evergreens

No AI required, though you can add some. You will.

For the tinkerers

Why these ship

Each project above has the three things the disk cleaner does not: a small blast radius, a target that stands still, and a finish line. A habit tracker that loses a habit costs you a streak. A cleaner that loses a file costs someone a home directory and earns itself a pinned apology post. PDFs will not rotate their cache format next month. PDFs have no roadmap. “Done by Sunday” is a claim no honest cleaner has ever made. Build the small thing, finish it, ship it.

The takeaway

Deciding what not to build is the same judgment that separates a useful tool from a liability. Pick the small thing, finish it, and ship it. The disk cleaner will still be there. Ideas are patient. Home directories are not.

Last reviewed