Final concept description

The EveryBot Scanner is a physical part identification tool built as an extension of EveryBot, an accessible robotics learning platform designed for middle and high school students. The scanner addresses a core barrier in robotics education. Not just cost and complexity, but the simple challenge of not knowing what a part is or what it does. Using a webcam connected to a computer and machine learning, students can show the scanner any part from the EveryBot system and instantly receive its name, a short description, and a link to further information.

Process and AI integration

The scanner was built by training a custom image classification model in Google's Teachable Machine, with approximately 1,200 samples per component class to ensure reliable identification. The model was then integrated into a p5.js interface that handles the live camera feed and displays results in real time, displaying the part name, a brief description, and a reference link. This pairing of a no-code ML training tool with a creative coding environment kept the build accessible and rapid to iterate on, which directly mirrors the philosophy of the larger EveryBot project.

Outcomes, prototypes, interactivity

The final prototype is a working, interactive scanner that identifies EveryBot components on demand. The system is fully functional. Beyond the classroom, the scanner has clear real-world utility in open-access spaces like makerspaces and school workshops, where electronics bins are shared and parts often arrive second-hand without labels or readable part numbers, especially motors. The EveryBot Scanner turns an unlabeled component into a learning moment.