This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What I Built
IRL Side Quest is a real-world adventure generator powered by locally running Gemma 4. It creates fun, bite-sized quests that encourage people to step away from their screens, explore their surroundings, and make ordinary moments more interesting.
Feeling bored? Generate a side quest! You might be challenged to find something unusual, discover a new walking route, observe nature, or complete a small outdoor activity.
The goal is simple: turn everyday life into an adventure, one quest at a time.
The app includes AI-generated quests, different quest categories, a simple XP and leveling system, achievements, and locally stored progress. Players can complete quests at their own pace without accounts, subscriptions, or complicated setup.
Code
IRL Side Quest is currently in the concept and planning stage. This submission presents the project's vision, proposed architecture, and planned use of locally running Gemma 4. The implementation and source code are not yet available and will be developed as the project progresses.
How I Built It
IRL Side Quest uses Gemma 4 through Ollama for local AI inference. The application sends quest preferences to a local Next.js API route, which communicates with Ollama and returns a structured quest for the interface to display.
Technology stack:
- Next.js and React — user interface and application routing.
- TypeScript — type-safe application logic.
- Tailwind CSS — styling and responsive design.
- Ollama + Gemma 4 — locally generated quests using an open-weight AI model.
- localStorage — saves player XP, completed quests, achievements, and progress in the browser.
- Lucide React — icons and visual elements.
The architecture avoids cloud AI APIs, API keys, authentication, and a dedicated database. Quest completion is self-reported, keeping the game simple and easy to build on.
Why Does Open Innovation Matter?
IRL Side Quest is designed around the idea that AI should empower people without demanding their attention or personal data.
Using an open-weight model through local inference means quest generation does not require sending prompts to a third-party AI service. Players can experiment with locally available models, inspect the architecture, and adapt the project to their needs.
There are also no per-request cloud AI charges. Once the model is installed, quest generation can work without an internet connection, provided the local application and model are running.
Most importantly, the app challenges the usual goal of keeping people engaged on a screen. Instead, AI creates the experience and then encourages the user to put the device away.
The ultimate objective isn't to spend more time in the app. It's to make something interesting happen outside it.
Prize Categories
Best Use of Gemma — IRL Side Quest uses the Gemma 4 open-weight model through Ollama to generate real-world side quests locally.
Top comments (0)