🔍 Inspiration

Doctors and therapists often take quick, messy notes — hard to read, hard to share, and hard for patients to understand. We wanted to create a tool that bridges that gap: one that understands scratch-style notes and turns them into clear, useful summaries for both patients and providers. Especially in mental health, tracking progress across sessions is key — but it's rarely done well. We built Echo to change that.

💡 What it does

Echo is an AI-powered tool that takes unstructured clinical notes (like those scribbled during therapy or doctor visits) and turns them into:

  • Patient-friendly explanations in plain language
  • Structured summaries for providers (for EMRs, handoffs, referrals)
  • A timeline of sessions showing changes and recurring issues over time

It works with both physical and psychological health notes.

🛠️ How we built it

We used:

  • Groq API (Llama) for language understanding and summarization
  • A custom note parser for shorthand & medical phrases
  • Lovable for frontend, Supabase for backend

🧱 Challenges we ran into

  • Parsing messy, shorthand notes that lack grammar or structure
  • Separating physical from psychological content accurately
  • Making summaries both medically accurate and patient-friendly
  • Avoiding hallucinations or oversimplification in sensitive cases

🏆 Accomplishments that we're proud of

  • Built a working prototype in less than 3 hours
  • Generated both patient and doctor summaries from examples
  • Simulated timeline view of therapy session notes
  • Maintained tone-sensitive output (empathy in mental health, clarity in physical care)

📚 What we learned

  • Medical AI needs both accuracy and emotional intelligence
  • Lovable can do backend and integrate with AI APIs with natural language

🚀 What's next for Echo

  • Add support for image-to-note input (real-time dictation)
  • Expand timeline analytics for long-term care tracking
  • Pilot with small clinics or mental health startups

Built With

Share this project:

Updates