Rephrase

Inspiration

As someone who feels communication is often misunderstood, and knowing full well how hard it is to start conversations, our app aims to remedy this issue with the help of artificial intelligence.

What it does

Our app unifies communication and takes pressure off of the user by allowing them to write their thoughts once, and have it reworded, or, "rephrased", into multiple formats and formalities. On top of that, our app helps people to expand their communication and enrich existing ones by utilising social media profiles to create compelling conversation starters.

How we built it

The mobile application is developed with React Native (TypeScript and JSX). Back-end APIs are also written in TypeScript. https://icon.kitchen was used to create the splash screen and mobile icon.

Challenges we ran into

API key limitations - running out of quota on multiple occasions during testing. Merging branch conflicts.

Accomplishments that we're proud of

As a team, none of us have any prior experience with React Native or web development, and so to make a fully functional mobile application and learn a lot about its architecture in just 24 hours is an impressive feat.

What we learned

React Native application structure such as functional programming concepts and state management to update the UI, TypeScript principles and how respective code is written, UI design and user experience improvement: implementing modern interface design.

What's next for Rephrase

To expand the scope of the project, we hope to allow deeper integration with social networks such as Instagram which would allow for improvements such as the ability to scour the user's following list and identify accounts that they do not actively engage with in order to proactively suggest conversation starters. Further, the app could be used as a translation tool between languages which translates but also provides localisation and context relevant to the situation for enhanced communication. This could be used in conjunction with formality tools already implemented so conversations are less likely to be misunderstood when working with language barriers.

In future, we plan to improve privacy of the application by using an on-device language model rather than sending potentially confidential user message contents to Gemini's API.

Built With

• apis
• gemini
• reactnative
• typescript