Inspiration

Exams are a test of both knowledge and mental strength. Many students know the answers, but due to stress, anxiety, fear, or panic, they perform poorly.

We discovered that current online exams and learning platforms only assess students' responses, not how they feel. This inspired us to create Exam Mind AI, an AI system that detects and supports a student's emotional state during exams or study sessions in real time.

Our goal is to make exams more equitable, healthy, and human-centered.

What it does

Exam Mind AI is a web-based, emotion-aware AI assistant for students. It: Listens to a student’s voice during an exam or study session Detects emotional states like Calm Mild stress High anxiety Panic Responds in real time with: Breathing guidance Focus and confidence reminders The system does not provide answers or help in cheating. Instead, it helps students regain focus and emotional balance so they can perform to their true potential.

Teachers and institutions also get an anonymous emotional analytics dashboard to understand overall student stress levels.

How we built it

Exam Mind AI was built by Google AI Studio by prompting the two mastered prompts and some other minor changing prompts. link For prompting, we use the claude.ai and Gemini Pro. Also, website deployment and hosting is done by the edgeone pages

Challenges we ran into

Emotion detection accuracy: Human emotions are complex, so we had to carefully tune thresholds to avoid false panic alerts. Real-time performance: Ensuring low latency so feedback feels natural during exams was challenging. Ethical concerns: We designed the system to be privacy-first and supportive, not intrusive or judgmental.

Accomplishments that we're proud of

Built a working real-time emotion-aware AI system Successfully integrated voice emotion detection in a web app Designed a calm and student-friendly UI Created an ethical AI solution focused on mental well-being Delivered a strong demo suitable for real exam scenarios

What we learned

Emotional intelligence is as important as technical intelligence in AI systems. Small, supportive interventions can make a significant difference. Building responsible AI requires more than just accuracy. Hackathons are about impact and not just technology.

What's next for Exam Mind AI

In the future, we plan to: Support multiple languages and accents Improve emotion detection using multimodal inputs Deploy in online exams, competitive tests, and remote learning platforms Partner with educational institutions for real-world pilots

Built With

  • edge-one-pages
  • gemini-pro
  • google-ai-studio
Share this project:

Updates