Price Is Right

Everyone loves a good deal. Point your Meta Ray-Ban glasses at an item and instantly know if it's a steal. Powered by Gemini Vision, with multi-source pricing across several APIs.

Comment

Inspiration

As a team, collectively we have had several experiences in thrift and record stores being up charged for things that could have been bought cheaper online or from select stores.

What it does or

The app figures out what item you are looking at, searches for price info about it online, and speaks it to you with text-to-speech.

How we built it

We mapped out the core parts of the project (Glasses, Vision AI, Query building, API search, TTS), coded the parts to work by themselves, and then connected the parts together.

Challenges we ran into

We ran into some issues connecting the backend to the Meta Ray-ban glasses themselves. Handling inconsistent outputs from the Vision model and mapping them into reliable search queries was also difficult. Latency across multiple API calls made it challenging to keep the experience feeling real-time.

Accomplishments that we're proud of

Being able to efficiently chain together the needed APIs in order to go from a picture of an item to a spoken description of its price range.

What we learned

As a team we learned how to link together multiple APIs into an efficient pipeline as well as designing around latency constraints. We also gained experience working with mobile-to-backend communication and audio streaming workflows, such as through Eleven Labs, Gemini, and XCode.

What's next for Price Is Right

Improve accuracy of item recognition and price estimation, reduce latency across the pipeline, and expanding the sources that we would be able to pull data from.

Built With

Share this project:

Updates