MoodMirror

Main
Past entries
Generating
Generated
Stored in archive
Stress hub

Submitting to track 2: Make Anything, But Make it YOU

Inspiration

Mental health tools are often overwhelming. I wanted to make something gentle, non judgemental, but also a little magical, something that rewards reflecting on your feelings with a beautiful personal creation. MoodMirror invites users to check in emotionally, not with pressure or analysis, but with creativity. It transforms a week of simple mood into a dreamy AI generated painting.

What it does

MoodMirror is a non-invasive mental wellness web app where users log their moods once a day using a short journal entry and am emoji picker for emotion. After 7 entries, the app reflects on the users emotional trend and turns it into a unique AI generated mood painting. It also saves past paintings, creating an emotional timeline. Finally, there is a stress hub users can use to relieve their stress, either by playing a game, receiving an uplifting message or watching a relaxing video.

How I built it

It was built with Streamlit, a python based framework that made it quick and easy to develop a clean, interactive web interface. For the AI generated paintings, I integrated the Hugging Face Inference API with the Stable Diffusion XL model to transform mood summaries into art. All mood entries and paintings are stored using Streamlits session state, keep the app self contained.

Challenges we ran into

I wanted to implement a message that comforted, motivated or celebrated you based on your mood the past week, but I could not find a single text generating API that I could run through Hugging Faces Inference Provider. The only option would have been to run the API on my own personal network, which would have been too much. I had to settle for a simple message that could work no matter the mood trend.

Accomplishments that I'm proud of

I am proud of building a tool that transforms emotional reflection into something approachable. MoodMirror combines journaling, mood tracking and AI generated art in a seamless browser based experience. Each users week creates a unique painting based on their emotional patterns, turning mental check ins into something to look forward to.

What I learned

Honestly? I learned to take a step back and rethink my plans. I initially was going to use the transformers python framework to analyze the users input and decipher their mood, but after hours of errors and debugging, I took a break and realized the user can just tell us their mood. After that little brain fart, development proceeded smoothly.

What's next for MoodMirror

Next, I would want to implement an AI text generator that comforts, motivates or celebrates me based off my weeks moods to make the app genuinely therapeutic. I would also want to bring MoodMirror to mobile as a standalone application, so that users can log their feelings every week and not have to depend on leaving the session running on their device to avoid losing their past paintings. Making the experience more portable and more uplifting would be my next step.