Drishya Singhal

Posted on Oct 8

NatureFit: Turn Going Outside Into an AI-Powered Nature Quest 🌿

#devchallenge #weekendchallenge #hf26challenge #ai

🌐 Try NatureFit live: https://drishya-code.github.io/NatureFit/

🌿 NatureFit: Turn Going Outside Into a Nature Quest

What if going outside wasn't just about counting steps or calories?

What if your workout could also make you slow down, notice what's around you, and actually interact with nature?

That's the idea behind NatureFit.

I built NatureFit as an AI-powered outdoor fitness experience that turns physical activity and nature discovery into small quests.

Instead of simply telling you to exercise, NatureFit gives you a reason to go outside, explore your surroundings, move your body, and complete a challenge.

🌱 The Idea

Most fitness apps focus heavily on numbers:

Steps
Calories
Distance
Workout duration

Those metrics are useful, but I wanted to experiment with something different.

What if nature itself became part of the workout?

NatureFit combines two things in every quest:

🌿 Nature + 🏃 Movement

A quest can ask you to discover or observe something around you while also completing a physical activity.

The goal is simple:

Go outside. Move. Notice something.

And turn it into a quest.

🎯 How NatureFit Works

The basic flow is:

Create Quest → Get Nature + Movement Tasks → Start → Complete Tasks → Provide Evidence → Earn XP → Build Your Journal

When creating a quest, the user can choose things such as their activity, duration, fitness level, and environment.

NatureFit then generates a nature-and-movement challenge.

For example, a quest might ask you to:

🌿 Find and observe something interesting in nature.

🏃 Complete a short movement activity.

📸 Provide the required evidence.

⭐ Earn XP after completing the tasks.

The idea is to make outdoor activity feel less like a checklist and more like a small adventure.

🤖 Where AI Comes In

AI is used to generate varied outdoor quests based on the user's selected preferences.

I didn't want AI to be the entire product. Instead, it has a specific job: helping create interesting and varied nature + movement challenges.

There is also an offline fallback.

If AI generation isn't available, NatureFit uses its ready-made quest library instead.

The application clearly tells the user when the fallback is being used rather than pretending that an AI-generated quest was created.

🔒 Keeping AI Honest

One thing I really wanted to avoid was making AI claims that the application can't actually support.

NatureFit does not pretend to automatically verify someone's:

GPS distance
Steps
Physical activity
Location

Photo evidence is currently presence-based rather than pretending that an AI model has understood everything in the image.

Movement is also self-reported.

I think being honest about what the technology can and cannot verify is more important than adding impressive-sounding AI claims.

🏆 XP Without Easy Farming

There was another problem I wanted to think about:

What stops someone from simply generating quests repeatedly and farming XP?

Generating a quest gives 0 XP.

Starting a quest gives 0 XP.

XP is awarded through actual task completion.

Completing a quest also gives its completion bonus only once.

During testing, I actually found a bug where quickly double-tapping the "Complete quest" button could award the completion reward twice.

That was fixed with a one-shot completion guard, and a regression test was added for it.

This was a good reminder that testing the actual user flow can reveal problems that aren't obvious from simply reading the code.

📱 Built as a PWA

NatureFit is built as a Progressive Web App.

It can be used from a browser and installed like an app on supported devices.

The application also has an offline experience with ready-made quests, so losing connectivity doesn't completely stop the experience.

The goal was to make NatureFit something that actually makes sense outside, rather than something that only works comfortably on a desktop.

🧪 Testing It Properly

I didn't want to stop at:

"It works on my machine."

I ran NatureFit through a full testing pass covering the main user journeys, AI behavior, PWA functionality, accessibility, and edge cases.

The final results:

Test Area	Result
Logic	24/24
AI path	39/39
End-to-end	40/40
PWA	20/20
Accessibility	17/17
Edge cases	13/13
Total	153/153

I also tested multiple viewport sizes, including mobile and desktop, and checked keyboard navigation, focus visibility, contrast, reduced-motion behavior, and layout issues.

The live GitHub Pages deployment was also tested after the final changes.

🛠️ Tech Stack

React
JavaScript
Vite
Progressive Web App
AI-powered quest generation
Local browser storage
Playwright
Node.js

🌍 Why I Built It

The motivation was pretty simple.

We already have plenty of apps telling us how many steps we took.

I wanted to build something that gave us a reason to actually go outside and pay attention to where we are.

Instead of:

"Walk for 30 minutes."

NatureFit can turn that into:

"Go outside, discover something interesting, move for a while, and complete your quest."

That small change makes the activity feel more intentional.

🚧 Current Limitations

NatureFit is still a project, so there are some things it doesn't do yet.

Currently:

There is no GPS or distance tracking.
Movement evidence is self-reported.
Photo verification is presence-based.
There is no cross-device synchronization.
AI quest generation requires connectivity.
Physical camera capture and iOS PWA installation still need testing on real devices.

I'd rather document these limitations than pretend the application can do something it currently can't.