Confirm Details
Home Page
Review Results

Inspiration

Recalled products can still end up at stores and people’s homes. Buyers often do not know a product has been recalled, and checking model numbers across different government databases can be slow and confusing. We built ReCallout to make that process much faster and easier.

What it does

ReCallout lets users scan or upload an image of a product and helps identify key information such as the brand, model number, and product type. It then checks recall information from trusted sources and tells the user whether the product may have been recalled. Our goal is to help people make safer decisions before buying or using secondhand products.

How we built it

We built ReCallout as a web application that combines computer vision, text recognition, AI, and recall-data searching. The system analyzes an uploaded image, extracts identifying information, and uses that information to search relevant recall records. We designed the interface to keep the process simple: upload, identify, check, and review the result.

Challenges we ran into

One of our biggest challenges was product identification. Labels can be blurry, model numbers can be small, and different products display information in different locations. We also had to deal with recall information being spread across multiple sources and formatted differently. Matching a real-world product to the correct recall record without creating false matches required careful filtering and verification.

Accomplishments that we're proud of

We are proud that we turned a complicated recall-checking process into a simple user experience. Instead of requiring users to manually search through databases, ReCallout can use information directly from a product image. We are also proud that the project solves a real safety problem while using AI for something practical rather than just adding AI as a feature.

What we learned

We learned how computer vision, OCR, APIs, and AI systems can work together inside one application. We also learned that reliable AI is not just about recognizing something correctly; it is also about verifying information, handling uncertainty, and giving users results they can trust.

What's next for ReCallout

Next, we want to improve product recognition accuracy, support more recall databases, and make the system more agentic so it can automatically choose the best search strategy when information is incomplete. We also want to add barcode scanning, marketplace integration, recall alerts, and stronger confidence scoring. Eventually, ReCallout could automatically warn users about recalled products while they are shopping online or in person.