Echo

Talk. Feel. Heal.

Comment

Inspiration

Loneliness and mental health struggles are at record highs, yet professional help remains inaccessible to many. We created Empathy Echo to provide immediate, judgment-free emotional support using AI—bridging the gap between isolation and human connection.

What it does

• 🎧 Listens actively: Users share feelings via text/voice • 💬 Validates emotions: AI responds with empathetic, personalized feedback • 🔗 Recommends resources: Curated articles, hotlines, and coping strategies • 📊 Tracks mood trends: Identifies emotional patterns over time

How we built it

• Backend: Python + Flask • AI: Google Gemini API (fine-tuned for emotional intelligence) • Frontend: Bootstrap 5 (responsive design) • Database: Firebase (for mood tracking)

Challenges we ran into

Challenge: Gemini sometimes generated overly generic responses. Fix: Implemented a two-step prompt system: • First analyzes emotion/sentiment • Then generates response based on detected emotion + user history

Accomplishments that we're proud of

Frankly as a terrible artist, I successfully designed an Intuitive UI – Created a clean, accessible chat interface using Bootstrap 5 that works seamlessly on mobile and desktop.

Moreover, I had the opportunity to validated real-world need, receiving positive feedback from testers who found the tool genuinely helpful for stress relief.

What we learned

Firstly, I learned more about AI’s strengths & limits – while Gemini excels at NLP, it requires careful prompting to avoid robotic or insensitive replies. Therefore, prompt engineering would be extremely important when dealing with LLMs.

Furthermore, mental health tech is fragile – small wording changes in AI responses can significantly impact user trust and comfort. Therefore, it is important to invest more in this field to combat such issues.

What's next for Echo

• Multilingual support (non-English speakers are underserved in mental health tech) • Crisis detection (auto-connect to human help if suicidal ideation detected) • Voice tone analysis (using Whisper API for added emotional context) • Deployment!

Share this project:

Updates