This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What I Built
GET OUT is a local AI app designed to help people spend less time on their screens and more time in the real world.
The idea is simple:
The best AI session is the shortest session.
Instead of keeping users inside a chatbot, GET OUT gives them one real-world mission and then gets out of the way.
The user chooses:
- ⏱️ Time: 10 minutes, 30 minutes, 1 hour, or 2+ hours
- ⚡ Energy: Low, Normal, Active, Social, Quiet, or Surprise Me
The local AI then generates exactly one physical mission based on those choices.
For example:
You have 30 minutes. Explore a familiar route you normally ignore. Find one unusual detail you've never noticed before and investigate it for a few minutes.
PHONE DOWN. GO.
There is no recommendation feed, endless conversation, or AI companion.
The flow is:
USER → 2 CHOICES → LOCAL AI → ONE MISSION → PHONE DOWN → GO OUTSIDE
GET OUT is for people who feel like they are spending too much time on their phone or computer and need a small push to step away.
The goal isn't to maximize AI engagement.
The goal is to make the user leave the app.
Demo
Demo: [Add your demo link here]
The demo shows the complete flow:
- Choose how much time you have
- Choose your energy level
- Get one AI-generated mission
- Put the phone down
- Go outside
Code
GitHub: [Add your GitHub repository link here]
The project is open source and designed to run locally.
How I Built It
GET OUT uses open-weight AI with local inference.
Tech Stack
- Qwen3 4B — open-weight language model
- Ollama — local AI inference
- FastAPI — backend API
- Python — application logic
- HTML/CSS/JavaScript — frontend
- pytest — automated testing
The architecture is intentionally simple:
User
↓
Web UI
↓
FastAPI
↓
Mission Generator
↓
Ollama
↓
Qwen3 4B
↓
Safety Validation
↓
Mission Quality Validation
↓
ONE MISSION
↓
PHONE DOWN
↓
REAL WORLD
The model receives the user's available time and energy level and generates a mission designed around those constraints.
The generated mission is then validated before being shown to the user.
I intentionally avoided building another AI chatbot.
There is no endless conversation, recommendation feed, or AI companion.
AI generates the mission. Then the user leaves the screen.
Why Does Open Innovation Matter?
GET OUT is built around local, open-weight AI because the project does not need a large cloud AI service to work.
Using open AI makes it possible to:
- Run the model locally
- Avoid sending user context to a third-party AI API
- Experiment directly with the model and prompts
- Run without a paid inference API
- Modify the AI behavior
- Experiment with different open-weight models
- Keep the project accessible to other developers
But there is also a bigger reason.
Most digital products are optimized for:
more engagement → more sessions → more screen time
GET OUT is designed around the opposite goal:
less AI usage → more real-world time
That creates an interesting product philosophy:
We built an AI assistant whose success is measured by how quickly you stop using it.
The AI is not the destination.
It is the push toward the destination.
Prize Categories
- Open-Source Contributor
Top comments (0)