Reparo
An agentic AI repair platform that turns a product image into a repair plan, a parts list, and a path back to working order.
1ST PLACE / HACK CANADA 2026
From a broken object to a repair plan
Illustrative demoScripted example: no photo analysis, live inventory, or checkout runs on this page.
- Assessment accuracy
- 90%+
- Less search time
- 70%
- Real-time listings
- 1,000+
About this project
Built an agentic AI repair platform that uses Gemini vision models to classify products and assess repairability with over 90% accuracy, then generates step-by-step repair plans, required tools, and cost estimates.
Integrated the Shopify Storefront API and SerpAPI to automate part matching and streamline the repair-to-checkout flow, reducing search time for replacement parts by 70% with access to over 1,000 real-time listings.
Won 1st place in the Reactiv Track ($5,000) at Hack Canada 2026, and was shortlisted for the Most Complex AI Hack award.
How it works
- PRODUCT IMAGEPhoto of the product
- GEMINI VISIONRepairability assessment
- REPAIR PLANSteps, tools, and cost
- PART SOURCINGShopify + SerpAPI
Built with
- Python
- Gemini API
- React
- Node.js
- Swift
- PyTorch