Locus of You

LocusOfYou: a diagnostic engine for your potential. We find the hidden friction holding you back and provide the psychological blueprint to achieve your goals.

Comment

LocusOfYou MVP: Project Summary & Roadmap

This document summarizes the current state of the LocusOfYou application design and outlines the remaining macro-level steps required to complete the Minimum Viable Product (MVP).

Inspiration

This project was born from personal experience. After losing my career in tech during the COVID-19 pandemic, I found myself delivering packages to make ends meet. Along the way, I lost a sense of purpose and struggled with the motivation to get back on track. I built LocusOfYou because it's the tool I wished I had—an intelligent, empathetic coach designed to understand the deep-seated reasons for inaction and help rebuild momentum, one small step at a time.

What it does

LocusOfYou is a new type of AI coach that acts as a "master diagnostician of human motivation." Instead of offering generic advice, it gets to the root cause of why you're stuck.

The user journey begins with a conversational "Dynamic Assessment" that determines your unique "Motivational DNA" and user archetype. This culminates in a free, personalized "Motivational Snapshot" that provides a powerful "Aha! Moment" of self-insight.

For premium users, the app provides continuous, adaptive coaching. When you're struggling with a goal, its core diagnostic engine uses a unique synthesis of Expectancy Theory, Self-Determination Theory, and Goal-Setting Theory (ET→SDT→GST) to pinpoint the underlying issue and prescribe a targeted "micro-victory" to get you back in motion.

How we built it

We chose a modern, composable architecture to focus on the unique AI logic rather than infrastructure overhead.

Frontend: A responsive React application built with Vite and styled with Tailwind CSS. We used bolt.new to rapidly scaffold and iterate on the UI components.

Backend: A serverless Cloudflare Worker provides the API layer, running the core application logic at the edge for low latency.

AI Engine: LangChain.js is the "brain" of the application, running inside the Cloudflare Worker to orchestrate the complex diagnostic chains and interact with LLMs via Cloudflare Workers AI.

Database & Auth: Supabase handles all user authentication, a PostgreSQL database (managed via D1), and cloud storage for user avatars.

Subscriptions: RevenueCat, integrated via its web SDK and a Supabase Edge Function webhook, manages all payments and user entitlements, providing a reliable "single source of truth".

Challenges we ran into

Designing a "Low-Friction" Assessment: Our biggest challenge was gathering deep psychological data without making the user feel like they were taking a tedious quiz. We solved this by designing the "Dynamic Assessment"—a single-screen, conversational flow with a progress bar, varied interactions (tappable cards, sliders), and a "skip ahead" option to respect the user's time.

Avoiding the "Template" Feel: The initial "Snapshot" screen, while clean, felt like a generic report. To make it feel truly personal, we moved away from uniform cards and evolved the design to use abstract data visualizations (like the "Spectrum Bar" for Personal Agency) and a narrative "My Observation" summary that tells a story with the user's data.

Crafting a Compelling Call to Action: We iterated multiple times on the final CTA. Vague buttons like "Continue" or "Activate Coach" felt weak. We discovered the key was to make the final screen a "Celebration & Bridge" that builds on the user's success, leading to the final, emotionally resonant CTA: "Keep My Momentum Going."

Accomplishments that we're proud of

The Diagnostic Engine: We're incredibly proud of creating an AI that goes beyond surface-level advice. Its ability to diagnose the specific cognitive belief (Expectancy, Instrumentality, Valence) behind a user's inaction is the core innovation that sets it apart.

An Empathetic & Non-Judgmental UI: We were meticulous about the "Language of Agency." We designed every screen, data visualization, and piece of copy to be supportive and non-judgmental, ensuring the user never feels they've received a "bad score," only a starting point for growth.

A Truly Seamless User Journey: We designed a cohesive flow that takes a user from a state of curiosity, through a valuable "Aha! Moment" of self-discovery, to a tangible "micro-victory," making the final subscription offer feel like a natural and logical next step.

An Ethical-First Framework: We firmly established the "Coach, Not a Therapist" boundary and designed a non-negotiable "Crisis Intervention Protocol" from the outset, ensuring the app is a responsible and safe tool for personal growth.

What we learned

The Power of First Principles: Constantly referring back to our core principles ("Diagnose, Then Prescribe," "Make the Invisible, Visible," etc.) was our north star. It allowed us to make clear, consistent decisions and resolve complex design challenges.

Iteration is Key to Quality: The final designs for the Snapshot and First Step screens are vastly superior to the initial versions. This was only possible through a rigorous process of critique, feedback, and willingness to redesign core components until they felt right.

Human-Centered AI is Everything: The most powerful AI is useless if the user doesn't trust it or feel understood by it. We learned that the most important work was in designing the conversation, the empathetic framing, and the non-judgmental feedback system that surrounds the AI's logic.

Built With

Share this project:

Updates