My family uses four laptops and five mobile phones, and all of them seem to be running out of storage space. We want to organise our family data and create reliable backups, but first I needed to understand what was consuming space on our laptops.
My own five year old laptop was nearly full. I wanted to recover or reserve around 100 GB for better organisation and backups.
I did not want to pay for another commercial cleaning product. I was also cautious about downloading unfamiliar cleanup applications because of concerns about advertisements, spam, misleading prompts, scams, privacy, and tools that might delete something important.
I therefore decided to build my own application. I wanted something local, transparent, simple, and immediately useful.
That real need became Laptop Cleaner.
Laptop Cleaner is designed for people who may not be technically confident, including families, students, and senior citizens. It helps users understand their storage before they decide what to do.
The Windows desktop application shows the largest files, recursive folder totals, storage usage by file type, a conservative Safe Review Plan, and a detailed HTML report.
Safety is the central design principle. Laptop Cleaner is completely read only. It cannot delete, move, rename, or copy personal files, and it never opens file contents. It does not claim that a file is safe to delete. Instead, it explains which areas may deserve careful human review.
ChatGPT helped me shape the idea, define the safety boundaries, and plan the user experience. Codex then helped turn those requirements into working code at remarkable speed. Going from a real household problem to a tested desktop application within a few focused hours felt amazing. Codex proved to be an incredibly practical coding tool.
During OpenAI Build Week, I used gpt-5.6-sol for a meaningful technical improvement. It helped implement recursive folder aggregation so folder totals include every nested subfolder. The design calculates totals during scanning, stores them efficiently in SQLite, prevents double counting, and keeps selected scan roots separate. It also helped strengthen synthetic tests and checks for destructive operations.
The Hackathon Edition was validated on the original problem laptop. It measured 315,903 files using 89.6 GB across the selected personal folders while confirming that no personal files were changed.
I built Laptop Cleaner because I needed it myself. It worked, I could use it immediately, and I was happy with the result. That was when I realised this practical Codex assisted build could also become my OpenAI Build Week submission.
Laptop Cleaner does not perform backups yet. It provides the safe storage understanding needed before a family decides what to archive, back up, or review.
Log in or sign up for Devpost to join the conversation.