Better Everything

AI Coaching Meets meditation app — Calm, accessible, and intentional - designed to help you make real progress. Think, decide, and work better.

Comment

Demo - Full version : https://www.ghazanfaralikhan.us/long-product-demo Demo: YT Version: https://youtu.be/H5qnxTWkJcc

the logout button is on the top right on the Profile Tab

Better Everything — Hackathon Submission Narrative

Inspiration

I am a big believer in coaching, mentoring, therapy, mindfulness, and deliberate practice — and I apply these principles in my own life daily.

For years, I’ve built personal systems to track growth, unlock new skills, set goals, build habits, and automate workflows. I believe growth should be continuous, measurable, and system-driven.

I invest heavily in self-improvement and continuously experiment with physiological and psychological techniques to improve my state, focus, and performance. I also listen and follow many coaches across a broad spectrum from life in general to more specific ones like financial or communication coaches.

I also love building products — especially 0→1 initiatives that bring real value to people’s lives. I’m passionate about product design, product management, and building in public. The idea of putting something directly into users’ hands to test and refine deeply motivates me.

This project is inspired by Simon’s @BetterCreating brief, Notion systems, YouTube videos, and brand — particularly the idea that creators can build powerful, leveraged systems for growth.

What It Does

Better Everything is an AI-native, multi-coach platform designed to provide advanced and specialized coaching agents that:

  • Learn about the user over time
  • Adapt to their values, pace, challenges, preferences, and biases
  • Deliver measurable outcomes using scientifically grounded frameworks and techniques

The platform provides free coaching agents across major life domains.

Users can:

  • Adjust personality, tone, and coaching style
  • Connect the system to their existing workflows
  • Build custom agents with full control over model, prompt, context, guardrails, and evaluations

The system also generates personalized artifacts such as:

  • Vision boards
  • Roadmaps
  • Affirmations
  • Concept breakdowns
  • Task lists
  • Calendar events and reminders
  • Animated explainers for topics explored with coaches

The goal is not just conversation — but transformation.

How I Built It

Better Everything was built using:

  • VibeCodeApp (primary build environment)
  • Claude Opus 4.5
  • Figma
  • Replit
  • Midjourney

The architecture evolved during development from a stateful Make.com + OpenAI flow to a cleaner, more deterministic agent orchestration model.

Challenges I Ran Into

This project was built under significant constraints.

  • I initially implemented a stateful onboarding and routing system in Make.com with OpenAI coach agents, but ran into complex routing and state-management issues.
  • I was building simultaneously in Replit and VibeCodeApp, comparing outputs in real time.
  • Three days in, I made a difficult but strategic decision to pivot — abandoning Make.com and Replit code and rebuilding exclusively in VibeCodeApp.
  • To make time for building, I reduced Uber driving hours and did not meet rental requirements, resulting in my vehicle being suspended.
  • I did not have funds to participate and had to borrow to make this possible.

Personal Challenges

  • I discovered the hackathon two weeks after it had already started.
  • I am in between jobs and homes and living with someone and I can mostly only work out of public places, libraries, cafes.
  • My Apple Developer account was blocked prior to the hackathon due to an issue with my card and it wasn't technically possible to get it reinstated in time for the hackathon, so I had to coordinate with a friend to use their account. This posed operational issues where I had to work with them for 2 factor authentications with a 6 hour difference (They're in the UK and I am in the US) and work with them to get the app launch on testflight and testing of payments.
  • On the final day, unstable WiFi disrupted server connections throughout the night, corrupting UI and flow updates without clear error signals. So I wasn't aware that my fixes were being corrupted and not applied. I lost six hours and over $50 in usage costs.
  • final few hours: I need to push the build with In App Purchases but I cannot start the submission process on Expo because I am using a friend's Apple Developer account and Expo is requesting pin code for 2-factor authentication which is sent to my friends account, who is fast asleep in the UK (its past 2am there now)

Despite this, I completed and shipped.

Accomplishments We’re Proud Of

  • A psychologically-informed onboarding flow designed to keep user motivation high
  • A product experience that leads to an early “aha” moment
  • Progressive disclosure to prevent overwhelm
  • An agent system that:
    • Extracts topics from conversation
    • Creates running sessions
    • Stays warm, human-like, and non-prescriptive
    • Generates session summaries, action items, interventions, and follow-ups
    • Rich user profile context capture (personality types, values, energy patterns, mental models, life history) *The system behaves more like a structured coaching engine than a chatbot.

What I Learned

  • Design systems should be modular and implemented separately
  • Guardrails and evaluation loops are critical in AI-native products
  • Planning architecture early prevents costly pivots
  • Prompting techniques dramatically affect output quality
  • Agent architecture design is a product discipline
  • Vibe coding accelerates iteration but requires structural discipline
  • Vibe designing helps align experience and emotion
  • Learnt about Mobbin, growth.design & BuiltForMars for UIUX
  • Learnt how to use VibeCodeApp & Replit
  • Learnt about RevenueCat *Learn about App Developer Program and how to use App Store Connect
  • Learnt how to use Figma
  • learnt Agent fundamentals and how Agentic systems work
  • Leant debugging

What’s Next for Better Everything

Next steps focus on validation, refinement, and growth:

  • Implement feedback loops with real users
  • Fine-tune UI/UX
  • Improve agent performance and reliability
  • Ship additional free coaching agents
  • Instrument analytics and KPIs
  • Run A/B tests on onboarding, design, and copy
  • Define funnel targets (acquisition, activation, engagement, retention)
  • Define financial targets (conversion, LTV, CAC, MRR)
  • Explore partnerships with human coaches
  • Develop a GTM strategy across X, YouTube, TikTok, blogs, influencers, and reviews
  • Build a roadmap informed by user feedback, tech advancements, and competitive analysis

The long-term vision is to build a scalable, AI-native coaching infrastructure that meaningfully improves human performance and well-being.

Built With

  • chatgpt
  • make.com
  • midjourney
  • replit
  • vibecodeapp
Share this project:

Updates