About This Project
Overview
BetterEveryday operationalizes the idea that small daily actions compound: users log reflections through tokenized check-in links, track skills and projects through their lifecycle, and get a weekly AI-generated 'brutal review' scored on depth, consistency, and variety of effort. A skill-gap engine cross-references check-ins, project updates, and optional GitHub/LeetCode activity to surface avoidance patterns — all running on Supabase Postgres with row-level security and GitHub Actions cron automation.
Motivation
I kept starting habit trackers and abandoning them within a week, so I built BetterEveryday to hold myself accountable with something a plain checklist app couldn't do: an AI that reviews my week honestly instead of just congratulating me for showing up.
Lessons Learned
Tuning the weekly review prompt to be genuinely useful rather than either falsely encouraging or needlessly harsh took several iterations — the 'brutal but fair' tone is much harder to hit consistently than either extreme.