Voice Lab

• 

  photo

• 

  Just a photo

Inspiration

Lab results scare seniors. We built a voice-first Labs Analyzer so grandparents with poor eyesight can hear—not squint at—their results in plain language, in their language (EN/FR/AR/VN).

What it does

Upload or speak about your lab PDF → AI extracts values, flags abnormalities, and speaks back a calm, jargon-free summary with clear next steps. No reading required.

How we built it

• PDF → Structured Data: Qwen-VL + OCR for multilingual lab parsing
• Clinical Logic: Rule-based severity engine (green/yellow/red flags)
• Voice UX: Qwen-Audio TTS with slow, clear speech + empathetic persona tuning
• Simplicity Layer: One-tap voice input, zero typing, large audio feedback

Challenges we ran into

• Decoding messy, scanned lab PDFs across 4 languages
• Translating medical ranges ("HbA1c 8.2%") into reassuring, actionable voice guidance
• Designing for low digital literacy: no menus, no jargon, no confusion

Accomplishments that we're proud of

• 90%+ accuracy flagging critical values while keeping explanations under 30 seconds
• Voice flow tested with seniors: "Finally, I understand my blood work"
• Fully offline-capable voice mode for low-connectivity users

What we learned

Healthcare accessibility isn't a feature—it's the foundation. When designing for seniors, clarity beats cleverness, and voice isn't optional, it's essential.

What's next for Labs Voice

• Pilot with Elfie's elderly user cohort for real-world voice UX validation
• Add family caregiver mode: "Share this summary with your daughter" via one-tap audio message
• Integrate with Elfie Care to auto-schedule follow-ups when critical values are detected

Built With

• qwen