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
Presenting iris- a mobile/tablet-first AI point-of-sale system built with Expo, React Native, Supabase, Roboflow, and OpenAI. It uses the device camera for real-time object detection and LLM-powered naming, letting stores build tickets or receive shipments automatically. No barcode scanner required.
The architecture is ready for cloud scaling, with clear paths to use AWS Rekognition, Lambda, DynamoDB, and S3 for enterprise-grade deployments.
**Because iris depends on mobile hardware, the web simulator can’t show the full flow, Expo Go is the preferred experience.**
I don't know why it posted the video twice, but it was 1:15am so I'm going to give myself a pass.
Other notes:
I tested an early version of this in my brick and mortar stores just for the memes- it was surprisingly flexible with a lot of different product types and variants....
The fun part about multi-item detection and CLIP embedding here is that we can use the user workflow of receiving new items to train a model and get consistent object detection improvements.
Thanks to Tempo, AWS & Contra for the challenge. This was the most fun one yet for me.
(2/2)
Incredible
Thanks Joe!
Brilliant!
Thank you much, Mike!
This is tremendous! Thank you for sharing, Robert.
Love this! Worked on POS system with Expo, React Native Last year and this would have been a good feature addition (Object detection and LLM-Powered naming)
great out put
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
