# Project export: Touch Grass

This document was generated by HackStack to give an AI agent context about a hackathon project. Sections are labeled with their provenance; content marked as truncated was cut to keep this document small.

## Project metadata

- Hackathon: UC Berkeley AI Hackathon 2026
- Tagline: Touch Grass is an AI-powered accountability app that helps users break out of doomscrolling on TikTok, Instagram, etc by delivering personalized interventions that keep them focused on their goals.
- Devpost: https://devpost.com/software/touch-grass-trsawn
- GitHub: not linked
- Team: contributor stats unavailable

## Devpost submission (written by the team)

### Inspiration

We noticed that most screen-time apps fail because they rely on restrictions, timers, and app blockers that are easy to ignore. When people open TikTok, Instagram, or Reddit, they're usually looking for a break, not trying to waste time. We wanted to build something that understands why someone is scrolling and helps them redirect that energy into activities they'll actually enjoy. Instead of forcing users off their phones, Touch Grass acts like an accountability partner that nudges them toward healthier choices.

### What it does

Touch Grass is an AI-powered digital wellness coach that helps users reduce mindless scrolling and stay focused throughout the day. When a user opens distracting apps like TikTok, Instagram, or Reddit, Touch Grass analyzes their habits, interests, current context, and upcoming commitments to generate personalized interventions. Instead of simply blocking apps, it redirects users toward activities they're likely to enjoy and follow through on.

### How we built it

We built Touch Grass as a full-stack AI application that combines behavioral tracking, retrieval-augmented generation (RAG), and personalized recommendations. User sessions, intervention history, and behavioral signals are stored in a structured database, while Ollama and Qdrant power our AI retrieval system. When a user starts scrolling on distracting apps, Touch Grass retrieves relevant past interventions and generates personalized recommendations based on their habits, interests, and previous successes. We also integrated Google Calendar to make suggestions context-aware, allowing the AI to consider upcoming classes, meetings, and deadlines when helping users stay on track.

### Challenges we ran into

One of our biggest challenges was creating interventions that felt genuinely helpful rather than annoying. We quickly realized that generic notifications were ineffective, so we had to design a system that learns from user preferences and previous successes. Another challenge was building a retrieval system that could efficiently surface relevant user history while keeping data isolated and personalized. Balancing personalization, privacy, and performance required several iterations of our database and retrieval architecture.

### Accomplishments we're proud of

We're proud that Touch Grass goes beyond traditional app blockers and creates personalized interventions tailored to each user. We built a system that remembers what activities users enjoy, tracks which suggestions they accept, and adapts over time. We're also proud of creating a full end-to-end AI experience that combines behavioral data, retrieval systems, and large language models into a practical tool that addresses a real-world problem.

### What we learned

Throughout this project, we learned a great deal about retrieval-augmented generation, vector databases, personalization systems, and behavioral design. We also gained experience designing AI systems that operate on structured user data rather than relying solely on conversational context. Most importantly, we learned that helping people change habits requires understanding motivation and context, not just enforcing restrictions.

### What's next

Next, we want to make Touch Grass even more proactive. We're exploring deeper calendar intelligence, smarter behavioral modeling, and real-world activity recommendations based on location, time, and user interests. We also plan to improve our intervention engine by learning from long-term user behavior and identifying which strategies are most effective for different users. Our vision is to build an AI companion that helps people stay intentional with their time and develop healthier digital habits.

## README (from the GitHub repository)

No README available.

## Detected evidence (automated analysis)

No repository was indexed for this project. Claimed technologies below could not be checked against code.
- JavaScript (language) — claimed on Devpost, not found in the code
- Ollama (technology) — claimed on Devpost, not found in the code
- PostgreSQL (technology) — claimed on Devpost, not found in the code
- Python (language) — claimed on Devpost, not found in the code
- React (technology) — claimed on Devpost, not found in the code
- SQL (language) — claimed on Devpost, not found in the code

## Codebase structure

No repository index available.

## Key source files

No repository index available; no source files included.