Valetudo AI

AI medical assistant giving fast, clear, and source-based answers to real health questions. Built for parents, patients, students, and 40+. Safe meds, dosages, chronic care — less panic, more clarity.

Comment

Inspiration

Every day, millions search the internet for health answers — and end up overwhelmed by misinformation, panic forums, or outdated blogs. This is especially risky for parents, older adults, and people with chronic conditions.

I set out to build a tool that provides fast, reliable, source-backed answers to medical questions. Not a replacement for doctors — but an assistant that helps users act confidently and responsibly.

What I Learned

  • Working with Perplexity’s Sonar API revealed how powerful real-time, cited knowledge can be.

  • Prompt design for medical use is both science and empathy — clarity and tone matter.

  • UX = trust in healthtech: calm colors, intuitive UI, and clean layout build user confidence.

How I Built It

I designed 7 prompt templates based on real user needs:

  1. Symptom-based advice
  2. Drug interaction checks
  3. Safety timing
  4. Safer treatments
  5. Compare effectiveness
  6. Recommended exercises
  7. Latest research

The user can also ask a general question without using templates, or use auxiliary filters, ask a clarifying question (follow-up).

Users input data in a friendly form, and the system fetches live answers via Sonar Pro, with citations.

The frontend was built with React + Tailwind, backend on Flask, and optimized for trust — accessible, and visually calm.

Challenges

  • Clarity without diagnosis: Messaging the tool as guidance, not a doctor.

  • Prompt reliability: Ensuring consistent, cited, helpful output from the model.

  • Filter interaction: Some filters limit the response scope when used together. I had to balance flexibility and relevance to avoid blank or weak results.

  • API budgeting: Using mocks + fallback logic to stay within Sonar quota.

What's next for the AI Medical Assistant

Valetudo AI currently empowers individual users — next is B2B expansion.

Future plans:

  • Integrate with healthcare providers

  • Embed into research & pharma platforms

  • Support medical education tools

All while staying true to its mission: clear, evidence-based answers without noise or panic.

Built With

Share this project:

Updates