This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What if the best thing AI could do for us wasn't keeping us engaged with another screen, but helping us put that screen away?
That question inspired WildPath AI, an open-source AI-powered outdoor adventure planner built for the Hacktoberfest 2026 Touch Grass challenge.
Instead of generating more content to consume, WildPath AI helps transform a simple idea into a real-world outdoor experience.
The mission is simple:
Spend less time planning your next adventure and more time actually experiencing it.
What I Built
WildPath AI is a responsive web application that helps users discover outdoor activities and create personalized adventure plans using AI.
Imagine finishing work and wanting to spend an hour outside.
Instead of searching through dozens of websites, you can tell WildPath AI:
"I have 60 minutes, want something peaceful, and would prefer an easy nature walk."
WildPath AI is designed to turn that request into an outdoor activity recommendation.
π² Key Features
1. AI Adventure Planner
A natural-language planning experience that considers outdoor preferences and helps users decide what to do next.
2. Outdoor Discovery
Explore curated outdoor destinations with useful details to help choose an activity.
3. Personalized Adventure Preferences
Choose activities based on interests, available time, and preferred difficulty.
4. Adventure Journal
Keep a record of outdoor experiences and revisit past adventures.
5. Responsive Web Experience
WildPath AI is designed for desktop and mobile browsers, making it easy to explore without installing a native application.
How does it help people touch grass?
WildPath AI is designed around a different relationship with technology.
Instead of encouraging endless engagement, the application has a simple goal:
Open the app β Find an adventure β Go outside.
The screen is a starting point, not the destination.
Demo
π Live Application
[WildPath AI β Web Demo]
https://wildpath-bcmigta89-nishant-sabbarwal.vercel.app/
π¬ Product Video
The video introduces the idea behind WildPath AI and demonstrates the intended experience of moving from digital planning to outdoor exploration.
For the technical demo, I also want to show the actual application interface and the AI-generated planning experience.
Code
WildPath AI is open source.
π» GitHub Repository:
https://github.com/Nishant3007/wildpath_ai
The repository contains the web application source code, including its frontend, adventure-planning interface, and AI integration.
Developers can explore the implementation, contribute improvements, or experiment with different models.
How I Built It
My goal was to build a project where open-source AI is part of the actual product architecture rather than simply a label attached to it.
Technology Stack
| Layer | Technology |
|---|---|
| Frontend | Next.js |
| Language | TypeScript |
| Styling | Tailwind CSS |
| AI | Browser-based open-weight model integration |
| AI Runtime | Transformers.js |
| Local Data | Browser storage |
| Deployment | Vercel |
| Version Control | GitHub |
1. Building the Web Experience
I chose Next.js to create a responsive application that could be deployed publicly and tested from a browser.
The interface focuses on simplicity, with nature-inspired visuals and minimal friction between deciding to go outside and taking action.
2. Integrating Open-Source AI
The AI planning architecture uses browser-based inference rather than requiring every user request to be sent to a proprietary language-model API.
This makes it possible to experiment with openly available models and gives developers more control over how recommendations are generated.
The initial version focuses on demonstrating the planning experience. Further improvements include stronger model-based personalization and more reliable geographic route data.
3. Building for the Real World
One important engineering principle was to separate AI-generated suggestions from factual geographic information.
A language model can help interpret preferences, but real-world routes, distances, and navigation need verified map data.
This separation is important for making the product genuinely useful beyond a prototype.
4. Deployment
I used GitHub for source control and Vercel for web deployment.
The result is a publicly accessible project that can be tested without installing a mobile application.
What's Next?
The roadmap includes:
- Offline trail maps and downloaded routes
- GPS-based adventure tracking
- On-device nature identification
- Improved local AI inference
- Installable Progressive Web App support
- More detailed adventure journaling
These are planned improvements rather than claims about the current MVP.
Why Does Open Innovation Matter?
For me, the most exciting part of building WildPath AI is the possibility of creating useful AI experiences without making every interaction dependent on a closed, paid API.
π 1. Freedom to Experiment
Open models allow developers to experiment with different inference approaches, model sizes, and planning behavior.
I can improve the system without rebuilding the entire application around one proprietary AI provider.
π 2. Privacy by Design
Outdoor applications can involve sensitive information, including location preferences and personal activity history.
Browser-based inference creates an opportunity to process user requests locally rather than automatically sending them to a remote AI service.
That doesn't mean every part of the application is automatically private or offline, but it provides a stronger foundation for building privacy-conscious features.
π° 3. Lower Dependence on Paid Inference
With compatible open-weight models, inference can run on a user's device.
This creates an opportunity to reduce recurring server-side inference costs, although model downloads and device resources still have practical costs.
π 4. Technology That Encourages Real-World Experiences
AI products often optimize for longer sessions and increased engagement.
I wanted to explore the opposite idea.
What if successful AI interaction meant the user closed the application and went outside?
WildPath AI uses AI to help people spend less time using AI.
That's the philosophy behind this project.
Where Open AI Makes a Difference
The most valuable advantage is architectural flexibility.
Instead of requiring a remote proprietary model for every planning request, WildPath AI can evolve toward local inference, model swapping, and user-controlled data.
The trade-offs are also real: browser compatibility, model download size, memory usage, and inference performance.
These challenges are part of what makes building with open AI interesting.
My Agent Session
I used AI-assisted development to help plan the product architecture, develop the web MVP, and iterate on the user experience.
The development process included:
- Translating the Touch Grass theme into a product concept
- Designing the adventure-planning experience
- Selecting a web-first architecture
- Integrating an open-source AI inference approach
- Preparing the project for GitHub and deployment
DevRelay Session: [Add session link here if recorded]
Prize Categories
Hacktoberfest Open-Source AI Challenge β Week 1: Touch Grass.
Final Thoughts
WildPath AI started with one simple observation:
We have powerful tools to explore almost anything digitally, but sometimes we need a little help remembering to explore what's around us physically.
I wanted to build something that makes that first step easier.
The goal isn't to replace outdoor experiences with AI.
It's to use AI to help create more of them.
Less scrolling. More exploring. πΏ
Built with β€οΈ for Hacktoberfest 2026.
#opensource #ai #nextjs #hacktoberfest
Top comments (0)