JungleDex: Turning My Childhood Pokémon Dream Into a Real-World Adventure 🌿

Ritik — Sun, 11 Oct 2026 13:03:29 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

What I Built

Growing up, I was a huge Pokémon fan. I loved watching Ash go on adventures, discover new Pokémon, explore different places, and learn about his discoveries through the Pokédex. Finding a rare Pokémon was always exciting, but what fascinated me most was the idea of going on an adventure and discovering something new.

When I came across the Touch Grass hackathon theme, I thought: what if I could bring that childhood feeling into the real world?

That's how JungleDex — The Living World Dex was born. 🌿

JungleDex turns exploring nature into a game-like adventure. Instead of searching for fictional Pokémon, users can discover real plants, insects, birds, animals, and interesting landscapes around them.

Here's what you can do with JungleDex:

AI-powered discovery: Capture a photo of a plant or another natural object and get identification and information about it.
Nature Radar: Uses nearby observations from iNaturalist to help predict what you might encounter in your surroundings.
Collectible discovery cards: Save your findings in your own Living World Dex and share them with friends.
Quests and progression: Explore through quests, track distance and steps, earn XP, and level up.
AI conversations: Use Gemma to have conversations about discoveries you've saved, ask follow-up questions, and learn more about the natural world.
Model choice: Switch between Gemma and Gemini for AI analysis.
Privacy and offline storage: Keep discovery data locally using IndexedDB and choose what you want to share.

I also discussed the idea with my wife, and she suggested adding quests to motivate users to go outside and explore. That became an important part of the experience.

My goal is simple: I want an ordinary walk to feel like a small adventure, where you never know what interesting thing you might discover next.

Demo

🌿 Try JungleDex:

jungledex-web.onrender.com

I've tested JungleDex on plants around my home. The AI returned identification details that matched my observations, and I was able to share the discovery cards through WhatsApp and Instagram.

Code

GitHub repository:

ritikkhndelwal-netizen / jungledex

JungleDex 🌿 — The Living World Dex

Turn real-world nature discoveries into holographic 3D collectible cards.
JungleDex is a gamified outdoor exploration platform that blends multimodal computer vision with tactile, retro handheld hardware aesthetics. Snap photos of plants, birds, insects, and scenic vistas to identify specimens, build your catalog, and level up your outdoor expeditions.

🌟 Key Features

📸 Multimodal AI Field Scanner

Instant Species & Terrain Identification: Point your camera at any organism or landscape to receive binomial classification, confidence metrics, and ecological field notes via Google Gemini Vision.
AI-Generation Detector: Inspects whether an uploaded nature image is a genuine camera photograph or a synthetic AI render (Midjourney / DALL-E / 3D model).

🎴 Living Field Dex & 3D Holographic Foil

Tactile Hologram Cards: Tilt your device or cursor to illuminate dynamic metallic foil sheens across Common, Uncommon, Rare, Epic, and Legendary rarities.
Export & Share…

View on GitHub

The project includes a React, Vite, and TypeScript frontend, a Node.js and Express backend, local IndexedDB storage, and integrations for AI-powered discovery and nature observations.

Feel free to explore the code, try the application, and contribute ideas or improvements!

How I Built It

I built JungleDex with React, Vite, TypeScript, Node.js, and Express. Firebase authentication and Firestore support are also part of the architecture, alongside IndexedDB for local discovery data and images.

For the nature radar, I integrated the iNaturalist API to use community observations to help predict possible nearby discoveries.

The AI architecture

I initially used the Gemini API for image identification and information. I also wanted to explore Gemma, but I was concerned about requiring users to download and configure a model locally for a browser-based application.

I therefore integrated Gemma through Google's AI Studio API for the conversational experience attached to saved discovery cards. Users can ask questions about their discoveries and continue learning after the initial identification.

I also added a model-selection switch so users can choose Gemma or Gemini for AI analysis.

This lets me explore two different AI experiences within the same application. Gemma is not just mentioned in the project description; it has a dedicated role in how users interact with their discoveries.

A few things I learned

One of the biggest design challenges was getting the UI to feel like a Pokédex. I wanted to recreate that feeling, and although I'm happy with the current result, I still want to improve it.

I also wanted discovery cards to be private as well as shareable. My first attempt could share only text. I then worked on storing card data and images in IndexedDB so the application could retrieve and format the complete card for sharing. The sharing process works, although it still takes some time.

These decisions taught me to think beyond the visible interface and consider privacy, data storage, and the complete user experience.

Why Does Open Innovation Matter?

For me, open innovation means that a project can grow beyond what one developer can build alone.

JungleDex started with a personal idea inspired by childhood memories. As I developed it, I learned more about AI integration, local storage, sharing, and designing an experience that encourages people to explore the real world.

Using Gemma gives me an opportunity to build a conversational learning experience around an open-weight model rather than relying exclusively on a closed model for every AI interaction. It also gives me room to experiment with how AI can help people learn about their surroundings.

I want JungleDex to become a project that other developers can explore, improve, and contribute to. There is plenty more I want to build, from better UI and additional challenges to a more fully local application.

I believe open innovation makes it easier for ideas like this to evolve through experimentation, collaboration, and community contributions.

Prize Categories

Confirm the categories you are entering against the official challenge rules before publishing.

Best Use of Gemma - include this if the challenge accepts your specific Gemma integration method.

Best Use of Render - JungleDex is deployed on Render, with the frontend and backend hosted as part of the application.

What's Next?

I want to keep improving the UI, introduce more quests and challenges, and explore regional and worldwide competitions that encourage people to discover more of the natural world.

I also want to explore turning JungleDex into a fully local application, reducing dependence on cloud services and internet connectivity.

I'm genuinely happy with what I've built and what I've learned. Seeing a childhood-inspired idea become a working application has been a rewarding experience.

If you're a Pokémon fan, I hope JungleDex brings back a little of that childhood excitement. If you enjoy exploring nature, I hope it gives you another reason to step outside and look more closely at what's around you.

Try it, share your feedback, and contribute if you'd like to help it grow.

Your next adventure might be closer than you think. 🌿

DEV Community: Ritik