Most people trying to fix a home appliance are stuck flipping through a PDF manual or guessing wh...Most people trying to fix a home appliance are stuck flipping through a PDF manual or guessing wh...
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Abdullah's avatar
Abdullah Amir
• 5h
Most people trying to fix a home appliance are stuck flipping through a PDF manual or guessing which part is causing an issue. OnSense set out to replace that with an app where entering a model number instantly surfaces specs, part breakdowns, and an AI assistant trained specifically on appliance repair knowledge — not a generic chatbot.
What I worked on
As part of the mobile engineering work on the React Native (Expo) app, I built the OCR module that lets users scan an appliance's nameplate with their camera to automatically identify the model and pull it into the system — removing the need to manually type in model numbers that are often worn, tiny, or hard to read.
Stack:
React Native (Expo) for iOS and Android
OCR/camera-based scanning for nameplate recognition
AI chat integration (OnSense GPT) for appliance-specific Q&A
Backend integration with the platform's parts and repair-knowledge database
What's in the app:
Model lookup by number, with detailed specs and part breakdowns
Exploded View module — an interactive, clickable diagram of appliance parts
OnSense GPT chat assistant for error codes, repair steps, and part compatibility
Nameplate scanning via OCR for fast, accurate model identification
Parts cross-referencing against verified compatibility databases
Backend DevelopmentMobile DevelopmentAndroidiOSExpo
Post image
Saad's avatar
Saad Zaman
• 5h
The OCR-on-nameplate angle is the right call, those plates are half worn off and users mistype the model constantly. Accuracy there lives or dies on preprocessing before OCR, glare and low contrast on metal wrecks raw recognition, so that is where the module earns its keep.
Abdullah's avatar
Abdullah Amir
• 4h
Rightly Said 👍
Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started