The problem Mood Builder solves

Why Mood Builder?

Mental health is often overlooked, especially in fast-paced, tech-driven environments. Thus, Mood Builder helps users to proactively manage their mental health through:

• Effortless Mood Journaling: Quick and intuitive logging of daily emotions and reflections. Users can either type their entries directly or upload journal documents, including handwritten files. The system leverages Document Parsing technology from the Upstage API to extract the journal from uploaded documents.
• Automatic Emotional Analysis: AI-powered interpretation of journal entries using the Upstage Solar Pro model to analyze user mood score, emotions, themes based on user input journaling.
• Add-on Features: Additional features include a comprehensive Mood Trends Dashboard with visual analytics and mood boosters, a convenient Recent Entries and Calendar View to review and visualize emotional patterns, and personalized Mood Boosters like guided meditations, breathing exercises, and gratitude practices.
• Chat with PinAI Agent: Chat with PinAI Agent is a highly empathetic and professional psychological counselor that leverages Upstage Solar Pro for real-time mood-enhancing suggestions. It automatically fetches user persona data from PinAI and references journaling entries and today’s mood of user from Supabase. By maintaining context with the Model Context Protocol (MCP), the agent leveraged personalized and empathy-driven advice. This synergy transforms raw journaling data into actionable insights that empower each user to improve their mental wellbeing.

Challenges I ran into

One of the main challenges I encountered while building this project was handling the processed data from the Upstage API. Since many of the features rely heavily on user-generated input (Journaling Data), the cron job needed to be carefully configured to process the API response, extract relevant information, and store it accurately in a PostgreSQL database hosted on Supabase.

Initially, I faced errors in the data handling process. To resolve this, I meticulously reviewed each step of the data flow and optimized the journal analysis prompts to ensure the API returned well-structured outputs. This made it much easier to parse and store the data reliably.

After these adjustments, the system was able to successfully process the response from document parsing, analyze the journaling data, and store all the relevant information in the database without issues.

Another minor challenge involved some UI inconsistencies that affected user experience. While I managed to implement several fixes, a few issues remain due to time constraints. I plan to continue improving the UI after the hackathon to ensure a smoother user experience.