Zave

• 

  Zave Home Screen

Zave | Universal Repair Intelligence

Zave is an advanced multimodal decision-support agent designed to demystify physical repairs. Built on the cutting-edge Gemini 3 architecture, Zave helps users navigate the classic “Fix vs. Replace” dilemma by providing instant diagnostics, real-time market cost comparisons, and interactive, hands-free coaching.

Inspiration

We were inspired by how often people throw away broken items simply because they don’t know how to fix them—or whether fixing them is even worth it. With rising costs, growing e-waste, and sustainability concerns, we wanted to empower people to make smarter, more informed decisions about repairing everyday objects instead of defaulting to buying new ones.

What It Does

Zave acts as a universal repair assistant that: Diagnoses physical issues using multimodal inputs Compares repair costs versus replacement costs in real time Guides users step-by-step through DIY repairs with hands-free interaction Helps users confidently decide whether fixing or replacing an item makes more sense The goal is to save money, reduce waste, and make repairs approachable for everyone.

How We Built It

We built Zave using Google AI and the Gemini 3 architecture, leveraging its multimodal reasoning and decision-making capabilities. Gemini 3 allowed us to: Interpret complex repair scenarios Generate actionable repair guidance Provide intelligent cost comparisons and recommendations The system was designed to prioritize clarity, usability, and real-world practicality.

Challenges We Ran Into

Our biggest challenge was ensuring working links, reliable outputs, and consistent responses across different scenarios. Integrating accurate external data and making sure results were presented clearly and correctly required significant iteration. Fine-tuning outputs so they were both technically correct and user-friendly was a major focus.

Accomplishments That We’re Proud Of Successfully building a functional repair intelligence agent on Gemini 3 Creating a tool that actively promotes sustainability and DIY culture Delivering clear, actionable repair guidance instead of vague suggestions Gaining hands-on experience with advanced AI implementation in a real product

What We Learned

We learned a tremendous amount about: AI system design and implementation The real power and flexibility of Gemini 3 Handling multimodal inputs and structured outputs Designing AI tools that solve practical, everyday problems This project deepened our understanding of how AI can bridge the gap between knowledge and action.

What’s Next for Zave

Next, we want to: Improve accuracy of diagnostics and cost estimates Expand support for more device categories and repair types Add deeper real-time market integrations Enhance hands-free and interactive repair coaching Scale Zave into a go-to platform for sustainable repair decisions

Built With

• google-ai-studio