QuitCoach: Micro‑Habit Companion for Student Smokers

QuitCoach helps student smokers quit one small step at a time, turning daily check‑ins and AI‑powered nudges into a personalized, judgment‑free path to a smoke‑free life.

Comment

Inspiration

Quitting smoking is one of the hardest challenges a person can face, and generic "cold turkey" apps often feel cold and robotic. We wanted to build something that feels like a human partner—a coach that doesn't just track numbers but understands the psychological struggle. QuitCoach AI was inspired by the idea that micro-habits and empathetic conversation can bridge the gap where willpower alone fails.

What it does

QuitCoach AI is an end-to-end smoking cessation assistant. Users input their current habits and their goal, and our custom Gemini-powered engine generates a personalized, day-by-day reduction plan. Beyond just a schedule, the app features a "Live Coach" chat interface that provides real-time support, motivational nudges, and personalized advice based on the user's progress and mood.

How we built it

We utilized FastAPI to create a high-performance backend, maintaining an in-memory state for rapid response times during the demo. The core intelligence is powered by Google Gemini 1.5 Flash, which we tuned to act as an empathetic cessation expert. The frontend was kept intentionally minimalist using Tailwind CSS to ensure the user stays focused on their journey without unnecessary distractions.

Challenges we ran into

One of the biggest hurdles was "Prompt Engineering" the AI to return data in a consistent format for our dashboard while maintaining a warm, human tone in the chat. We overcame this by implementing a dual-service logic: one focused on structured data extraction for plans and another on natural language processing for the coaching interface.

Accomplishments that we're proud of

We are incredibly proud of the seamless integration between the data-driven plan and the conversational AI. Seeing the "Coach" reference specific daily targets during a chat session makes the experience feel truly personalized and impactful.

What we learned

We learned how to leverage LLMs for both structured logistical tasks (like habit planning) and unstructured emotional support (coaching) simultaneously. We also gained deep experience in building stateful applications with FastAPI.

What's next for QuitCoach: Micro‑Habit Companion for Student Smokers

We plan to integrate a full RAG (Retrieval-Augmented Generation) pipeline to connect users with peer-reviewed medical studies and local support groups, turning the app from a simple coach into a comprehensive healthcare companion.

Built With

Share this project:

Updates