01
Problem
Team practice is structured; individual development at home mostly isn't. Coaches can't easily assign work, see whether it happened, or respond to it.
02
Approach
Build the loop end to end: coach assigns a week, player trains and films, coach reviews. Then test it with real teams and iterate on what actually limits usage.
03
Design decisions
- Training laid out as a weekly path so players always know the next step
- Short, fixed-length clips so review is fast for coaches
- Quick-feedback chips alongside written notes, to lower the effort of responding
- Safety defaults throughout because most users are minors
04
Technical build
- Flutter app on iOS, with Google Play approval in progress; Next.js marketing site and signup on Vercel
- Supabase Postgres with 20+ Edge Functions for checkout, uploads, moderation, notifications and scheduled cleanup
- Cloudflare Stream for video with short-lived authorized playback; clips auto-deleted after six months
- AI plan generation via the Claude API, with every drafted drill validated against a schema before a coach sees it
- Stripe subscriptions plus Apple and Google purchase verification
05
Outcome
NextTouch 2.0 is live with 3 youth teams and 75+ downloads/users (October 2026), expanding toward 10+ teams and a goal of 20+ teams by Fall 2027.
06
Lessons
- The binding constraint was coach consistency, not player interest. Automating reminders and weekly communication mattered more than adding features.