Spark: Habit Tracker

Most habit apps give you a blank checklist. Spark uses OpenAI to generate a personalized 3-phase roadmap from a single dream statement.

Comment

Inspiration

Gabby Beckford has built a community of ambitious women who dream big — but dreaming and doing are two different things. We noticed a critical gap: most productivity apps treat goals as checklists, ignoring the emotional journey of transformation. A woman who says "I want to run a marathon" doesn't need a to-do list; she needs a roadmap that meets her where she is and grows with her.

We were inspired by three things:

  • Gabby's PacksLight philosophy — travel light through life, but pack powerful habits. Our design system is literally named PaxLight as a tribute to this ethos.
  • The science of progressive overload — just as athletes build strength in phases, lasting habits require a Foundation → Growth → Mastery arc.
  • The celebration deficit — women often minimize their wins. We wanted every small victory to feel like a moment, not just a checkbox.

That's why we created Spark — and its companion, a glowing firefly mascot who guides you from your first dream to daily mastery. Because every spark can start a fire.

What It Does

Spark is an AI-powered habit tracker designed for ambitious women (late 20s–late 40s) that bridges the Inspiration → Action gap.

The Spark Firefly — Your Personal Companion

At the heart of the app lives Spark, an animated firefly mascot who accompanies users throughout their entire journey. Spark floats, glows, bounces, and celebrates with you — reacting to your mood, your achievements, and your profile. When you complete a habit, Spark lights up with a golden glow and confetti. When you're onboarding, Spark introduces itself: "Hi there! I'm Spark." It's not just a UI element; it's an emotional companion designed to make the habit-building journey feel less lonely and more joyful.

Deep Psychological Profiling (Backed by ~100 Research Articles)

Before showing a single habit, Spark runs a 24-question psychological assessment across 5 scientifically-grounded dimensions. We analyzed nearly 100 academic articles on habit psychology, behavioral science, and motivation theory to design this scoring system:

Dimension What It Measures Why It Matters
Motivation Clarity Goal precision, driving force, timeline realism People with specific catalysts succeed 3x more
Self-Discipline Follow-through, routines, procrastination patterns Predicts habit sustainability
Past Experience Previous habit attempts, app usage history, obstacles Identifies repeat failure patterns
Time & Energy Available daily time, energy levels, current wellness Prevents overcommitment
Support System Social sharing, accountability preferences Matches intervention to personality

Each dimension scores 0–100, producing a Spark Score (0–500) that classifies users into personalized profiles:

  • 🔥 Habit Champion (400-500) — 4 habits, 14-day foundation
  • ⚡ Rising Spark (300-399) — 3 habits, 21-day foundation
  • 🌱 Growing Seed (200-299) — 2 habits, 28-day foundation
  • 🌿 Fresh Start (100-199) — 1 habit, 38-day foundation
  • 🫧 Gentle Beginning (0-99) — 1 habit, 49-day foundation

This means a "Gentle Beginning" user gets 49 days of foundation phase with just 1 habit, while a "Habit Champion" gets 14 days with 4 habits. The app literally adapts its difficulty to who you are. We even detect specific obstacles — if you're forgetful, you get extra reminders; if you lose motivation quickly, celebrations are amplified; if you're time-poor, habits are capped at 10-minute micro-actions.

Core Experience

  1. Set Your Dream — Type a goal like "I want to write a book" or "I want to get fit for my 30s."
  2. AI Roadmap Generation — OpenAI generates a personalized 3-phase plan:
    • Phase 1: Foundation (~33%) — Easy wins, low friction (e.g., 10-min journaling)
    • Phase 2: Growth (~33%) — Build intensity (e.g., 30-min writing sessions)
    • Phase 3: Mastery (~34%) — Full capacity (e.g., daily chapter drafts)
  3. Daily Command Center — All habits aggregated by time of day with one-tap completion.
  4. Celebrations — Confetti overlays, "SHE DID THAT!" moments, streak badges at 7, 14, 30, 60, 90, and 365 days — with Spark the firefly bouncing and glowing alongside you.

Key Features

  • Power Moves — Daily micro-challenges to push comfort zones
  • Mantras & Mindset — AI-generated motivational content personalized to the user's goals
  • Discover Feed — Tips, success stories, and a daily community question
  • Question of the Day — Community-driven engagement with likes and answers
  • Calendar View — Visual completion heatmap across weeks and months
  • Journey Roadmap — Interactive timeline showing progress across all 3 phases
  • Smart Notifications — Habit reminders timed to when you said you feel most energized

Monetization (RevenueCat)

Feature Free Premium ($4.99/mo)
Habit tracking 3 habits max Unlimited
Basic streaks & calendar ✓ ✓
Discover content ✓ ✓
AI-powered roadmaps — ✓
AI coaching & mantras — ✓
Advanced analytics — ✓
Exclusive challenges — ✓

The free tier lets users prove the app works for them. When they hit the 3-habit limit and want more, they're already invested — creating a natural, low-friction upgrade moment.

How We Built It

Architecture: Clean Architecture + MVVM

  • Frontend: Flutter (Dart) — single codebase for iOS and Android
  • State Management: Riverpod (40+ providers across 12 feature modules)
  • Navigation: GoRouter with declarative routing and deep linking
  • Backend: Node.js + Express + TypeScript with Prisma ORM (PostgreSQL)
  • Database: Cloud Firestore (mobile, offline-first) + PostgreSQL (API)
  • Auth: Firebase Authentication (Email, Google, Apple Sign-In, Anonymous)
  • AI: OpenAI GPT-4o for 3-phase habit plan generation
  • Subscriptions: RevenueCat SDK with native PaywallView UI
  • Notifications: Firebase Cloud Messaging + local notifications

Design System: PaxLight

The entire UI is built on a custom design token system named PaxLight — a deliberate nod to Gabby's PacksLight brand:

  • Coral — Primary action, energy, confidence
  • Teal — Adventure, depth, exploration
  • Gold — Celebration, achievements, milestones
  • Typography: Playfair Display (bold headlines), DM Sans (clean body), Cormorant Garamond (inspiring quotes)

Psychological Engine:

The 24-question onboarding survey was built by analyzing ~100 research articles spanning:

  • Habit loop theory (Duhigg, Clear)
  • Self-determination theory (Deci & Ryan)
  • Implementation intentions (Gollwitzer)
  • Behavioral activation and motivation science
  • Procrastination psychology and temporal discounting

The scoring service uses weighted multi-dimensional analysis with special logic: reversed scoring for negative indicators (procrastination), capped scores to prevent inflation (wellness activities max at 30 points), and negative-to-positive resets for users who self-identify as consistent.

Backend API: 10 modular endpoints covering auth, goals, habits, completions, streaks, today, calendar, discover, users, and AI — all with JWT authentication, Zod validation, and comprehensive error handling.

Challenges We Ran Into

  1. Designing a fair psychological scoring system — Balancing the 5 dimensions so that no single answer could skew the entire profile required extensive iteration. Multi-select questions needed special logic: obstacle questions use negative scoring (each obstacle subtracts points), but users who select "I usually stick with habits" get a full positive reset. We tested dozens of weight combinations.

  2. RevenueCat + App Store Connect integration — Configuring in-app purchase products, linking them to RevenueCat offerings, and ensuring the native PaywallView rendered correctly on iOS required careful coordination between three dashboards (Xcode, App Store Connect, RevenueCat).

  3. AI response consistency — Getting GPT-4o to reliably produce exactly 3 phases with valid JSON structure required iterative prompt engineering. We added server-side validation with Zod schemas and automatic duration adjustment to guarantee phase durations sum correctly.

  4. Making the mascot feel alive — Spark the firefly needed to feel like a companion, not a sticker. We implemented 5 different animation states (floating, glowing, bouncing, celebrating, thinking), particle effects, dual-glow systems (gold + coral), and haptic feedback with sound effects — all while keeping the widget tree performant.

  5. Adaptive difficulty from survey to habits — Translating a 0–500 Spark Score into actual habit parameters (count, duration, foundation days, reminder timing) required mapping psychological dimensions to concrete UX decisions. A "Gentle Beginning" user sees a fundamentally different app than a "Habit Champion."

Accomplishments That We're Proud Of

  • The Psychological Profiling Engine — A 24-question, 5-dimension assessment backed by ~100 research articles that genuinely adapts the app experience to each user. This isn't a personality quiz — it's a behavioral science tool that determines habit count, phase duration, reminder timing, and obstacle-specific interventions.

  • Spark the Firefly — A mascot that users actually care about. The animated firefly with 5 mood states, particle effects, celebration bounces, and sound effects creates genuine emotional connection. It introduces itself by name, celebrates your wins, and glows brighter as you progress.

  • The 3-Phase System — No other habit app on the market generates progressive, phased roadmaps from a single dream statement. This is genuinely novel and directly addresses the "I don't know where to start" problem Gabby's audience faces.

  • PaxLight Design System — A complete, production-quality design token system (colors, typography, spacing, radius, shadows) named as a tribute to Gabby's PacksLight brand. It gives the app a distinctive, premium feel — far from generic "template UI."

  • "SHE DID THAT!" Celebrations — The confetti overlay with haptic feedback, Spark bouncing with joy, and gold glow effects when completing habits creates an emotional reward loop that makes tracking feel joyful, not tedious.

  • Full RevenueCat Integration — A complete freemium monetization stack with native paywall, subscription state management, feature gating, restore purchases, and real-time status updates across the entire app.

What We Learned

  • Personalization isn't optional — it's the product. The difference between a user who gets 1 habit for 49 days vs. 4 habits for 14 days is the difference between retention and churn. One-size-fits-all habit apps fail because people aren't one-size-fits-all.

  • Emotional design drives retention more than features. Our testing showed that Spark the firefly and the "SHE DID THAT!" celebrations were the #1 reason people wanted to keep using the app. The confetti matters more than the calendar.

  • Research-backed design takes time but pays off. Analyzing ~100 articles to build the scoring system was the longest pre-development phase. But it meant every survey question has a purpose, every score has scientific grounding, and every personalization decision is defensible.

  • RevenueCat simplifies IAP dramatically — but the App Store Connect side still requires patience. The SDK abstraction is excellent; the Apple review process is its own challenge.

  • AI-generated plans need guardrails. GPT-4o is powerful but non-deterministic. Wrapping AI output in strict Zod validation schemas was essential for production reliability.

What's Next for Spark: Habit Tracker

Short-Term (Post-Hackathon)

  • App Store Launch — Complete Apple review and publish to the App Store
  • Android Release — Flutter codebase is cross-platform ready; Google Play submission
  • Social Features — Accountability partners, shared goals, community challenges
  • Widget Support — iOS home screen widget showing today's habits and streak count

Medium-Term

  • AI Coaching v2 — Weekly AI-generated progress reports with personalized suggestions based on completion patterns and Spark Score evolution
  • Habit Stacking — Visual habit chain builder based on "Atomic Habits" methodology
  • Integration with Gabby's Content — Pull in PacksLight blog posts, podcast episodes, and travel challenges directly into the Discover feed
  • Apple Watch / WearOS — Quick habit completion from the wrist with Spark mascot complications

Long-Term Vision

  • Creator Platform — Let influencers like Gabby create and distribute custom habit roadmaps to their communities
  • Enterprise / Group Plans — Teams, book clubs, and accountability groups with shared dashboards
  • Localization — Spanish, French, Portuguese to reach Gabby's international audience
  • Spark Score Evolution — Re-assess users quarterly, showing how their psychological profile has grown

Spark: Because every big dream starts with one small habit.

Built With

Share this project:

Updates