ReflectAI

• 

  This is the interface of my web

• 

  I ask question by sending this picture that what about her personality said.

Inspiration

Many people want to improve themselves but don’t know where to start or feel uncomfortable sharing personal struggles. I wanted to build a private, supportive AI companion that helps users reflect on themselves in real time—without judgment, without social pressure, and in their own language.

As a student interested in AI and human-centered technology, I was inspired to create a tool that combines live camera input with ethical AI feedback to encourage self-growth and confidence.

What it does

ReflectAI is a real-time, AI-powered self-improvement application that uses live camera detection to understand visual cues such as posture, attentiveness, and presence. Based on these signals, the AI provides instant, constructive feedback focused on personal growth, confidence, and habit-building.

The app supports multiple global languages, so users receive feedback in the language they select. All responses are designed to be positive, ethical, and non-judgmental. No images or videos are stored, ensuring privacy and trust.

How we built it

The application uses live camera input processed locally for real-time detection. Key visual signals are converted into lightweight data, which is then sent to the Gemini 3 API for fast AI reasoning and personalized responses.

The frontend handles camera access, language selection, and real-time UI updates. The backend (optional) manages AI requests, performance optimization, and response formatting to ensure low latency and smooth interaction.

Challenges we ran into

One of the biggest challenges was reducing response latency while using live camera input. To solve this, we optimized detection frequency, minimized payload size, and structured prompts to prioritize fast, concise AI responses.

Another challenge was ensuring ethical feedback. We carefully designed prompts to avoid negative judgments and focused instead on encouragement and actionable self-improvement suggestions.

What we learned

Through this project, we learned how to integrate real-time inputs with AI reasoning, design multilingual user experiences, and build responsible AI systems that prioritize user privacy and well-being.

This hackathon helped us understand how AI can be used not just to analyze, but to empower people.

What’s next for ReflectAI

Future improvements include emotion trend tracking (without storing images), accessibility-focused features, more global language support, and optional personalized growth plans based on long-term user goals. ration

Built With

• css3
• gemini-3-api
• github
• google-ai-studio
• html5
• react
• typescript
• vite
• web-camera-apis