ReActive

App

Inspiration

In today’s busy world, many people struggle with mental fatigue and stress due to the overwhelming number of tasks and responsibilities they have to manage. I wanted to create a tool that not only boosts productivity but also promotes mental well-being by reducing the cognitive load associated with planning and organizing. Inspired by the idea of using technology to streamline daily routines, I developed Reactive—an AI-powered task assistant that transforms spoken thoughts into actionable to-do lists, helping users maintain mental clarity and focus on what truly matters.

What it does

Reactive is an AI-powered app that converts user speech transcripts into structured to-do list items, complete with interactive links for ease of use and productivity. Users can simply speak their tasks, and Reactive will automatically identify the actions, contacts, and time details, generating tasks that are easy to manage. The app categorizes tasks by type, such as email, SMS, calls, meetings, and visits, and creates interactive links for each task. With chronological sorting, time conversions, and location recognition, Reactive ensures that users have a clear, organized list of tasks ready for action.

How we built it

We built Reactive using Next.js with TypeScript for a robust and scalable front-end, ensuring a smooth user experience. The app integrates the Web Speech Recognition API to capture and transcribe user speech, which is then processed by Gemini AI to identify tasks, contacts, and other details. We used the Contact Picker API to enable seamless contact selection for creating interactive links. The app's architecture is designed with React for a component-based approach, and we styled it with Tailwind CSS for a modern, responsive interface. Throughout development, we focused on creating a solution that is both intuitive and effective in enhancing user productivity and mental well-being.

Challenges we ran into

Since the Contact Picker API isn't supported on Desktop, had to commit minor changes to github, test on mobile, reiterate Prompt engineering the AI service for the correct task/response format was something too

Accomplishments that we're proud of

Realtime transcription AI assist with streaming support Using the Contact Picker API to Inline links for calls, SMS, Email

What we learned

There was a lot with this project, specifically with its limited timeframe, had to move through a lot, but prompt engineering takes the spotlight － how to prompt AI to get what you, took a lot of back-and-forths but I got here