LiftFire - GymTracker

LiftFire turns workouts into a social promise—squad pledges, XP & streaks—while AI auto-progresses your weights. Show up, press Start, level up together.

Comment

Inspiration

I’m in the “I hate the gym” camp. I’ll start strong on Monday and ghost by Thursday. What finally clicked was treating the gym like a promise to friends, not a fight with willpower. If my squad is waiting, I show up. If it feels like a game, I keep going.

What it does

LiftFire turns workouts into a social commitment with a light game loop. Small squads make weekly pledges, every session earns XP and streaks, and friendly challenges keep momentum. An AI helper auto-progresses your weights (and suggests deloads) so “what should I lift?” never blocks you. Show up, press Start, level up together

How we built it

Mobile-first React + Vite + TypeScript, Supabase (Auth/Postgres/RLS) for users and data, and IndexedDB for offline sets. We used a PWA + service worker for instant launch and caching. The AI auto-progression is a rules engine that tracks recent performance and proposes next weights with guardrails. We worked spec-first (tiny end-to-end slices: DB → API → UI → test) to ship fast without breaking flow

Challenges we ran into

Keeping “social pressure” motivating without feeling punitive was tricky; wording and toggles mattered. Offline sync + streak logic led to edge cases. Service worker caching bit us (extension/unsupported schemes). And auth deep-links/redirects needed extra care to feel seamless on mobile. Issues with AI when the app turned more robust for sure, as having to take care of optimization with more hooks than at start. Then I have to create a sport-tracker manager backend app just to manage exercises, default workouts, challenges, goals, etc.

Accomplishments that we're proud of

We made “showing up” the default: one-tap start, auto-weights, visible streaks. Squads/pledges feel lightweight yet sticky. The offline experience is genuinely fast. And the spec-first workflow let us ship a cohesive slice end-to-end in days, not weeks.

What we learned

Social commitment beats solo willpower—copy tone matters as much as features. Guardrailed AI suggestions reduce decision fatigue more than we expected. Offline-first is worth the complexity; perceived speed = real retention. Tiny specs keep everyone aligned and shipping.

What's next for LiftFire - GymTracker

Integrate a real social feed with updates of all crew members Squad Mode v2: smarter nudges and “quick win” quests when someone falls behind. Challenges + Win Cards: shareable weekly summaries to spark referrals. AI v2: personalize progressions by exercise family and recovery signals. Admin tools: curate starter templates/programs and moderate public challenges.

Built With

Share this project:

Updates