Kill Sin Project

Overview

An AI-powered accountability app designed to help users confront and overcome destructive habits through daily reflection, structure, and grace.

Problem

Many habit and accountability apps optimize for streaks, shame, or surface-level motivation.
For people struggling with deeply personal behaviors, these tools often fail to provide context, compassion, or sustained engagement.

Kill Sin was created to address a harder problem:
How do you help someone consistently confront personal failures without discouraging them — and without removing responsibility?

My Role

• Concepted the product from first principles

• Designed the core product philosophy and UX language

• Built the front-end experience and state architecture

• Designed and trained AI conversation flows

• Implemented persistence, tracking, and progression systems

• Iterated continuously based on real daily usage (including my own)

System / Solution

Kill Sin is built as a daily-use system, not a one-time app.

Core components:

• AI-guided daily check-ins (morning & night)

• Context-aware accountability conversations

• Calendar-based win/loss tracking

• Streaks, records, and achievement systems

• Modular “opponent” framework for different habits

• Persistent memory to personalize responses over time

Tech & tools:

• React Native + Expo

• Node/Express backend

• Fine-tuned GPT models for conversational logic

• Async storage & structured state management

• Modular component architecture to support future expansion

Key design decisions:

• Framed accountability as training, not punishment

• Balanced structure with grace to prevent burnout

• Prioritized daily frictionless use over feature density

Outcome

• A fully functional mobile app in active development

• Multiple complete product surfaces (dashboard, logs, stats, achievements)

• Working AI conversation systems with memory and context

• Real daily usage and iterative refinement

• Clear path to beta launch

What I Learned

• Accountability systems fail when they rely on guilt instead of clarity

• Daily products live or die by friction, not features

• AI is most effective when constrained by clear philosophy and rules

• Designing for consistency matters more than designing for intensity