Inspiration
Emotional loneliness and silent stress are becoming increasingly common, especially among students and young adults. Many people hesitate to open up due to fear of judgment or not being understood. I wanted to create a space where users could express their emotions freely, without being analyzed, fixed, or judged.
InnerBloom was inspired by the idea that sometimes people don’t need solutions — they just need to feel heard. I wanted to explore how modern AI, when guided responsibly, can offer emotional presence and comfort rather than advice or diagnosis.
What it does
InnerBloom is an emotion-aware AI support companion that listens with empathy. When a user shares how they feel, InnerBloom gently reflects their emotions, validates their experience, and responds with calm, supportive language. The system is intentionally designed to avoid medical or psychological advice, focusing purely on emotional support.
How we built it
Google Gemini 3 (Gemini Flash) Google AI Studio Prompt Engineering Emotion-Aware Conversational Design
Challenges we ran into
One of the biggest challenges was defining clear boundaries for the AI — making sure it supports users emotionally without crossing into diagnosis or therapy. Designing prompts that balance empathy, safety, and responsibility required careful iteration.
Another challenge was deployment decisions, where I learned that not all AI apps require backend servers. Understanding the correct use of AI Studio versus Cloud Run helped me simplify the architecture and focus on user experience.
Accomplishments that we're proud of
Built an emotionally intelligent AI companion that prioritizes empathy over advice, creating a truly non-judgmental space for users. Successfully designed safe and responsible prompt boundaries, ensuring the AI offers emotional support without crossing into therapy or diagnosis. Leveraged Google Gemini 3 (Gemini Flash) to generate calm, human-like responses that adapt to users’ emotional states. Deployed the project as a public, accessible AI Studio app, removing friction for first-time users. Transformed a simple idea into a meaningful, human-centered AI product within a limited timeframe.
What we learned
Through this project, I learned how powerful prompt engineering can be when combined with advanced models like Gemini 3. I also gained insights into responsible AI design, especially for sensitive use cases such as emotional well-being.
This project strengthened my understanding of building meaningful, human-centered AI systems rather than purely technical ones.
What's next for innerBloom
Introduce mood tracking so users can reflect on emotional patterns over time. Add daily gentle check-ins and affirmations to build a sense of continuity and care. Support multiple languages, making InnerBloom accessible to more people. Improve emotional understanding using advanced emotion classification and context memory. Design privacy-first personalization, allowing users to feel seen without storing sensitive data. Expand InnerBloom into a mobile app for a calmer, distraction-free experience.
Log in or sign up for Devpost to join the conversation.