Skip to Main Content

Tintico Team

Built at Meta Global AI Developer Hackathon · Oct 26, 2026 · Remote

Tintico Team — Demo video

Every year we throw away 62 million tonnes of electronics, and less than a quarter gets recycled (UN Global E-waste Monitor 2024). A lot of those things could be fixed or built. People just get stuck on a confusing manual, make a mistake on step 3 and only find out on step 6. Chatbots don't solve this: they can tell you how, but they can't see what you did. StepSight watches you do it. You point your phone camera at what you're building or fixing, and Timber, a beaver guide, helps you step by step: - It finds the manual for your exact model. It can guess the model from a photo, and if the manual isn't online, you can photograph the paper booklet. - It splits the manual into small steps, one action each. - It lights up the right pieces on your own table (SAM 3.1 + Muse Spark) and shows a drawing of the step (Muse Image). - It checks a photo of every step (Muse Spark). If something is wrong, it tells you what to fix right away. - It's hands-free: you can ask "where's the long black axle?" mid-step (Muse Voice Transcribe). Safety is part of the design. Electric repairs start with an "it's unplugged" photo check that can't be skipped, and dangerous jobs like gas leaks or car brakes are sent to a pro. It works in English and Spanish, and I tested it with repeatable labs: 57/57 conversations understood, 96/96 step checks and voice commands right, 4/4 dangerous jobs sent to a pro. Try it: https://stepsight.vercel.app

Team