Dish-covery

Imagine a tool that can remove the stress and burden combing through a menu and not understanding what's best for you? With Dish-covery, you can translate, visualize, and explore menus in real time!

Comment

About Dish-covery

Ordering under pressure has never been easy. From having friends finding their exact order within a millisecond, to not understanding what's actually on the menu during client dinners. With Dish-covery, you can say goodbye ordering obstacles.

What is Dish-covery?

Dishcovery is a tool aimed to simplify the ordering process at restaurants. From our convenient web app, you can upload a photo of any menu in any language for Dish-covery to process. Dish-covery will then lift the text off the menu, breaking down each dish into its own card with information. Images of the dish will be displayed within these cards to visualize exactly what you will be ordering.

Our Purpose

Dish-covery is not just a tool to help with ordering. We designed and built Dish-covery with an emphasis on accessibility. Menu's have always been a struggle for those who dyslexia, language barriers, visual impairments, or aphantasia.

Some features that encapsulate these key considerations include:

  • Large visual pictures to showcase sample images of the dish to users
  • Speech input connected to an OpenAI chat bot to ask more about specific dishes
  • Menu translation to ensure users can understand what exactly is contained on the menu

How Dish-covery was made

Utilizing a react and tailwind front end, we created a clean and simple web app to greet users initially. All text and image inputs are translate through OpenAI implementation for text response generation, speech-to-text, and text-to-speech. Image upload and text identification was implemented through the Replicate backend. Text is then uploaded and stored through a mongodb custom database before being sent back to the user side.

Future Possibilities

To enable easier use in the future, porting the project to a mobile app for easier use is intended. As well, support for more translation languages and higher accuracy of non-latin scripts is considered for higher accuracy and easy of use. This could then be packaged and distributed to restaurants that use ordering tablets to provide accessible ways for their customers to try and enjoy their meals.

Built With

Share this project:

Updates