Used Claude to make a screen time app that let's me exclude certain apps

DaanyaalSobani published TrueScreenTime, a screen-time app built with Claude and hosted at github.com/DaanyaalSobani/truescreentime. The project’s pitch is simple: track device usage while letting the user exclude chosen apps from the totals. It appeared on Hacker News (item 49161351) under a Claude/Codex/Fable thread with 1 point and 1 comment at the time of capture.

The product mechanic that matters is selective exclusion. Stock screen-time tools often fold every foreground app into one number, so work tools, messaging, or always-on utilities inflate “usage” and bury the habits someone actually wants to change. TrueScreenTime treats exclusion as a first-class control: the user decides which packages count toward the report and which do not, so the metric tracks intentional focus rather than raw uptime.

For engineers and builders, the interesting signal is not a new OS API but how Claude was used to ship a personal utility that fills a gap in platform defaults. Screen-time data lives behind permission gates, usage stats APIs, and OS-specific foreground tracking; getting exclusion right means defining a clear inclusion set, persisting user choices, and keeping the reported total honest when apps are added or removed from that set. A single-author, AI-assisted repo that solves one sharp UX complaint is a concrete pattern for agent-assisted product work.

Competitively, Apple Screen Time, Digital Wellbeing, and third-party focus apps already own the category, but most optimize for limits, schedules, and blanket blocking rather than tunable denominators. TrueScreenTime’s differentiation is narrower: cleaner accounting of “time that counts.” On HN it had only 1 point and 1 comment at capture, so it is early and unproven as a market play; its value right now is as a reference implementation and a public demo of Claude-built tooling in a crowded, permission-heavy space.

Watch the GitHub repo for how exclusion rules are stored, whether the app stays single-platform, and whether the HN discussion grows beyond the single comment. If you build similar tools, copy the scope discipline: one measurable complaint (inflated screen-time totals), one control (app exclusion), and an open repo so others can audit the accounting logic rather than trust a black-box dashboard.