Cue: Your AI Coach

Home Screen
Your Coaches Screen
Coach Chat
Notifications/Cue Controls

Inspiration

Simon’s brief articulated a problem I’d been thinking about for a while.

Coaching works. But access doesn’t.

Professional coaching costs $200–500/hour. AI tools like ChatGPT are powerful but require non-technical users to become prompt engineers. There’s a gap between expensive human coaches and generic AI responses.

When Simon described his audience: people “obsessed with productivity, great design, and building systems”, I knew exactly who I was building for. These are people who understand the value of coaching but haven’t found an accessible way to get it.

Cue bridges that gap.

What it does

Open Cue, tap 'Mind', and tell your coach you're struggling with decision fatigue. Unlike ChatGPT, your coach already knows your goals, values, and past conversations. It asks one perfect question.

4 specialized AI coaches across core life areas:

Mind: Mental clarity, focus, decision-making
Body: Fitness, nutrition, sleep, physical health
Spirit: Purpose, relationships, mindfulness
Wealth: Financial planning, career growth, money mindset

And Custom Coaches based on your goals and challenges.

Core features:

Personalized goal setting for each life area with AI-refined objectives
Context-aware conversations that remember your values and challenges
AI-generated daily notifications (“cues”) for timely motivation
Custom coach creation for any life area
Freemium model: 15 messages/day free, unlimited with premium
Response time: Sub-2 second AI response time

How I built it

Frontend

React Native + Expo (TypeScript)
Zustand for state management
React Query for server state
Custom theme system with Libre Baskerville typography

Backend

Supabase (PostgreSQL, Auth, Edge Functions)
Row Level Security for data protection
Local-first architecture with server reconciliation

AI Integration

OpenAI API via Supabase Edge Functions
Custom system prompts for each coach specialty
Context injection for personalized responses

Monetization

RevenueCat for subscription management
Three tiers: monthly ($9.99), annual ($59.99), lifetime ($149.99)

Notifications

Expo Notifications with push tokens
Server-side cron jobs for scheduled delivery
AI-generated content personalized to user goals

Design

Warm cream (#F5F1EB) and forest green (#1B4332) palette
Libre Baskerville serif for headlines, system fonts for body
Minimal, premium aesthetic matching Simon’s brand

Challenges I ran into

1. Crafting AI Prompts That Actually Coach

Generic AI responses don’t cut it for coaching. Users don’t need “here are 10 tips.” They need a coach who asks the right questions and guides them to their own insights. Crafting system prompts that produce genuinely helpful, contextual coaching responses required extensive iteration.

2. Achieving Premium UX Without a Design Team

Simon’s audience expects exceptional design. As a solo developer, achieving that “beautiful, minimal, premium” feel required studying apps like Notion and premium wellness apps. Every interaction had to feel intentional.

3. Local-First Architecture Complexity

Making the app feel instant while keeping data synced with Supabase involved solving real technical challenges: optimistic updates, conflict resolution, RLS policy compliance, offline-first patterns.

4. Shipping Production Quality in 4 Weeks

Solo developer. Full-featured app. Auth, chat, goals, notifications, subscriptions, custom coaches. Every feature had to earn its place.

Accomplishments that I am proud of

Built something I'd actually pay for. After coding the MVP, I opened Cue and had a genuine coaching conversation about a decision I'd been avoiding for months.

Working MVP with 4 AI core coaches + custom coaches, real-time chat, goal setting, notifications, and subscriptions
Coaching that feels genuine — conversations & AI tailored notifications that lead somewhere, not generic chatbot responses
Premium design that matches Simon’s aesthetic: warm, minimal, sophisticated
Clean architecture that a team could maintain: proper TypeScript, Zustand stores, React Query patterns
RevenueCat integration production ready end-to-end with receipt validation and entitlement checking

What I learned

RevenueCat makes subscriptions manageable — the SDK handles iOS/Android differences elegantly
Supabase Edge Functions are powerful — serverless AI integration without managing infrastructure
UX matters more than features — Simon’s audience expects polish; a working feature isn’t enough if it doesn’t feel right