EchoGuard

"Behind every message, there’s a feeling."

Comment

🧠 Inspiration Mental health often goes unnoticed — especially online, where people share how they feel but no one truly “hears” them.
I wanted to create something that could listen without judgment, detect emotional patterns, and gently raise a flag if someone might be in crisis. During a late-night chat with a friend who said “I don’t think I can do this anymore,” I realized how powerful it would be if tech could act as a first responder. That moment inspired EchoGuard.

💬 What it does EchoGuard analyzes the emotions behind messages/text and checks for distress signals or crisis phrases.

  • Takes user input (text)
  • Detects emotional tone (happy, sad, angry, anxious, or neutral)
  • Checks for dangerous phrases (e.g. “I want to die”, “can’t go on”)
  • Stores emotional history locally
  • If a crisis is detected, it simulates an emergency alert It’s fast, light, privacy-respecting, and ready to plug into larger wellness platforms.

🛠️ How I built it

  • Backend: Python + Flask + Flask-CORS
  • Frontend: HTML, CSS, JavaScript (vanilla)
  • Data store: JSON file (for mock database)
  • Alert system: Crisis alert is simulated via console (can be extended with Twilio or email APIs)
  • Deployed: Vercel The entire app runs on lightweight infrastructure with zero external dependencies unless extended.

🧗 Challenges I ran into

  • Making emotion detection meaningful without using large transformer models (due to time and memory limits)
  • Designing a simple and welcoming UI that doesn’t overwhelm users
  • Handling edge cases in text (e.g. sarcasm or vague language)
  • Getting Flask and frontend to work smoothly across local and web hosting setups
  • Balancing between a "serious" tone and a friendly experience

🏅 Accomplishments I’m proud of

  • Built a full-stack emotional intelligence tool solo in a short time
  • Crisis detection logic works reliably for common phrases
  • Clean API structure that could be reused in mobile or voice-based apps
  • UI is cute, usable, and simple — without external libraries
  • Created something that actually feels helpful and meaningful ❤️

-📚 What I learned

  • Flask API structuring for emotion-based applications
  • Better text processing logic using keyword context
  • How to simulate critical alerting systems
  • The power of minimal UI design — keep it human, not heavy
  • That sometimes the simplest tools can make the biggest impact

🚀 What’s next for EchoGuard

  • Integrate OpenAI Whisper for voice-to-text emotional analysis
  • Add Twilio or email alerts for real-time emergency contacts
  • Build a mobile app interface using React Native or Flutter
  • Connect with licensed mental health resources via APIs
  • Allow users to track emotional trends over time with analytics
  • custom avatars, and supportive chatbot responses

Built With

Share this project:

Updates