Inspiration

When we arrived at college, everything hit us at once; there were new people, new places, new challenges. In the middle of all that chaos, we realized something: students rarely get the space to pause and actually process their emotions and thoughts. Journaling is powerful, but most of us don’t have the time or discipline to write essays about our feelings every day.

That’s why we built sine. Not a journaling app, but a personal reflection engine. With a single check-in—text, voice, or video, sine transforms your thoughts into AI-powered insights: bite-sized summaries, actionable advice, and motivational prompts you can actually use. Beyond that, it visualizes your personal journey with analytics on positivity, stress, goals, and mindset over time.

sine isn’t about writing more. It’s about reflecting smarter, thriving daily, and taking control of your life.

What it does

An AI-powered super-journal, sine helps you improve two major areas of your life.

Mental Wellbeing

With an easily accessible web interface, sine users like you can gain all the benefits of journaling on the go! For "video call" entries, our sophisticated (speech-to-text) STT model is able to filter out your voice from the business around you--in a crowded room, on the subway, or walking home from classes. After you put your thoughts into our platform, your entries are then fed into a large-language model (LLM), specialized to provide you with what you need—a stranger to rant to, a therapist, a career coach--the sky's the limit when it comes to sine! From the transcript of your entry, we provide you with an automated analysis of your current mood, stress, and progress, and how these factors evolve over time. Additionally, if you so choose, sine will give you advice and tips on how to improve your wellbeing—encompassing mental, emotional, and social health.

Productivity

Success is a two-sided coin, made of productivity and wellbeing. Without one, it's very hard to achieve the other. Not to fear, sine is here to help with any and all things productivity! Our LLM also analyzes your entries for relevant goals and tasks that you do which help you progress towards them,

How we built it

After we drafted our outlines we got to work drafting our code. We started with the STT model using whisper.cpp, then got to work on the frontend and backend using React.js. Using Open, AI, Github Copilot, and Claude to help us with syntax for our code, we focused on page-by-page implementation to create a smooth program whose pages interface well with each other.

Challenges we ran into

  • The graphs for data visualization took a while to appear on the webpage because the frontend and the backend weren't initially as smooth as they are now.
  • Surprisingly, formatting the JSON gave us a challenge, because it took some time for us to figure out that the "JSON" we got from Gemini's response to the user's rant was not an actual JSON. It was a string. Once we figured that step out, we could convert the string into a JSON and then parse through it to create a nicely formatted text output. ## Accomplishments that we're proud of
  • Our STT model is very promising. It can recognize the voice of someone and prioritizes the closest person (if there are multiple people in the background), allowing people to use our app from busy spaces where they might be overwhelmed and in need of guidance.
  • We also learned how to collaborate as a group, as most of us had only met each other within the week, or even the day of! ## What we learned
  • Teamwork and delegation is crucial to getting the job done and staying productive and on track
  • More about how LLMs work with each other to create a functional code flow
  • How to plan for what we could conceivably achieve in 36 hours based on what we wanted to do ## What's next for sine
  • Optical Character Recognition - for people who do like journaling on paper, but still want to reflect with a buddy
  • Facial Emotion Recognition from Video - to personalize feedback even more because tones don't always convey true feelings
  • Optimize whisper.cpp - get faster responses and analysis
  • Video Call - have the app respond to you in real time, with real voice, AI-generated responses to add accessibility and depth

Built With

Share this project:

Updates