Trash Gremlin

Inspiration

We were inspired by the simple but powerful idea that cleaning up our communities could be fun.
We saw litter in our parks and on our streets and wondered: what if we could turn the chore of picking up trash into a rewarding game?

Combining our love for “catch ’em all” monster-collecting games with our passion for the environment, the concept of Trash Gremlin was born.
We wanted to create an experience where every piece of litter spotted and cleaned contributes to a larger, more engaging goal — nurturing and collecting unique digital creatures.

What It Does

Trash Gremlin is a mobile app that gamifies environmental cleanup.
Users explore their real-world surroundings to find and scan litter. The app features two main modes:

• Feed Mode: Users can “feed” their existing Gremlins by scanning litter, helping their creatures level up and grow stronger.
  
• Capture Mode: Cleaning up litter gives users a chance to capture new, unique Gremlins to add to their collection.
  

Beyond the core loop, the app includes a social platform where users can add friends, view Gremlin collections, and compete on leaderboards to see who’s making the biggest impact — turning a solitary task into a community-driven mission.

How We Built It

We designed the user journey and interface mockups in Figma, focusing on creating an intuitive and beginner-friendly experience.
The app was prototyped using React Native for cross-platform compatibility, with a lightweight backend powered by Node.js.

The core of our project is the AI-powered trash detection.
For our proof-of-concept, we successfully integrated the Trash Detection API from Roboflow Universe, allowing us to identify trash type and analyze its condition to provide three key metrics for our game:

1. Waste Category
   The AI model identifies the litter material (e.g., plastic, paper, metal, glass), forming the foundation for all further metrics.

2. Recyclability Index
   A score (0–100) indicating how easily the item can be recycled.
   Based on standards like DIN EN 45555, it evaluates:

   • Material composition
     
   • Ease of disassembly (if applicable)
     
   • Compatibility with recycling infrastructure
     For example, a plastic bottle would score higher than a complex, multi-material package.
     

3. Carbon Footprint Potential
   An estimate of the litter’s CO₂ equivalent (CO₂e) impact.
   Using methodologies from the EPA and GHG Protocol, it multiplies the waste weight by a material-specific emission factor depending on its likely end-of-life (e.g., landfill, incineration).

Challenges We Ran Into

Our biggest challenge came during the UI/UX design phase.
We envisioned high-fidelity 3D Gremlin models to make them feel alive and unique — but our 3D generation tool, Lovable, struggled to consistently maintain visual style. This led to a lack of cohesion across Gremlins.

On the technical side, the AI integration is not yet fully live.
While the “Manus AI thinking...” animation currently serves as a placeholder, we successfully validated the full Roboflow API pipeline, setting the stage for real-time AI analysis in the next build.

Accomplishments We’re Proud Of

• Building a complete gamified loop that turns litter cleanup into an engaging, rewarding experience.
  
• Validating our AI proof-of-concept with Roboflow — confirming we can accurately identify trash and generate actionable environmental metrics (Waste Category, Recyclability Index, Carbon Footprint Potential).
  
• Establishing the core gameplay system that merges education, environmental action, and fun.
  

What We Learned

We learned the importance of rapid prototyping and leveraging external APIs to accelerate innovation.
Using a robust pre-trained model like Roboflow’s allowed us to implement advanced AI features without building an ML pipeline from scratch.

We also learned to prioritize consistency over ambition.
While 3D modeling was an exciting goal, we realized that a cohesive, polished 2D experience provides a far stronger foundation for future growth.

What’s Next for Trash Gremlin

Our next steps include:

1. Full AI Integration:
   Connecting the Roboflow Trash Detection API directly to the live camera feed for real-time AI analysis.

2. 3D Gremlin Pipeline:
   Revisiting our 3D asset workflow to create consistent, expressive Gremlins using improved generation tools.

3. Social & Community Expansion:
   Introducing new Gremlin species, community cleanup events, and competitive leaderboards to keep players engaged.

Ultimately, we want Trash Gremlin to grow into a global movement — where picking up litter feels as exciting as catching your next favorite creature.

References

[1] ComplyMarket. (2024). How to Calculate Recyclability of Your Products.
[2] Emission Index. (2024). How to Calculate Carbon Emissions From Waste.

Built With

• lovable
• manus
• trashdetectionapi