# Docarion

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Godswill Olemgbe](https://cerebralvalley.ai/u/dr_gee)
- **GitHub:** https://github.com/mannasync/docarion-meta-multimodal
- **Demo video:** https://www.youtube.com/watch?v=-dXTo4Eb28M
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/468

FOUNDING TEAM (TEAM #468 — DOCARION):
- Dr. Godswill Olemgbe, MBBS — Founder & CEO (@evidencechecked)
- Dr. Ayotomiwa Ayilola, MBBS — Co-Founder & CMO (@silverpenydr)
- Yusuf Alani Olayode — Co-Founder & CTO (@Doctware)
- Divya — Co-Founder & Docarion Central Admin (@KunnujiC)

Docarion is a clinical intelligence and telehealth platform built for frontline healthcare across Sub-Saharan Africa—connecting community pharmacies (PPMVs), rural clinics, care homes, labs, and hospitals with licensed physicians powered by Meta Open Source AI.


WHAT WAS BUILT DURING THIS HACKATHON (STARTUP TRACK DISCLOSURE)

While Docarion provides the underlying multi-facility EMR foundation, the entire Meta AI clinical intelligence layer shown in our demo video was built from scratch during this hackathon across three core pillars:

1. Adaptive AI History-Taking & Longitudinal Fixed-History Engine (Meta Muse Spark + SAM 3.1):

- Built the concise 3-turn adaptive plain-English and Nigerian Pidgin/Hausa/Yoruba/Igbo symptom intake engine (ai-intake.service.ts, intake-history.service.ts) with a strict zero-leak patient safety boundary.

- Built the Once-Only Longitudinal Fixed History Editor (FixedHistoryEditor.tsx) that permanently saves a patient's allergies, medications, past medical/surgical, family, and social history so returning patients are never re-asked static questions.

- Built 2-hour incomplete intake session resumption and Two-Phase Doctor Notifications (alerting on-duty physicians when history starts and when the case is ready to claim).

- Integrated self-hosted Meta SAM 3.1 on Modal T4 GPUs into the intake and Nurse Care Station flows to scan Malaria RDT strips, measure chronic wound surface area (cm²), redact patient headers from paper lab slips, and verify NAFDAC medication packs.


2. Clinical Diagnosis, Differential (DDx) & Auto-SOAP Decision Support System (Meta Muse Wonder):

- Built the Pre-Claim Physician Brief (PhysicianBrief.tsx) on the telehealth queue so doctors can review the patient's synthesized HPI, red-flag checks, and allergies in 15 seconds before claiming.

- Built the Live In-Chat Clinical Co-Pilot Card (live-chat-ai-copilot-card) inside the consultation room, powered by Meta Muse Wonder, which continuously reads the intake history, nurse vitals, and live chat to generate: (a) Ranked ICD-10 Differential Diagnoses (DDx) with 1-click adoption, (b) On-the-spot red-flag & follow-up questions with 1-click "Ask in Chat", (c) NAFDAC-verified treatment & WHO AWaRe antibiotic stewardship suggestions, and (d) Auto-synthesized 4-part SOAP notes (S, O, A, P).

- Built the Flexible e-Prescription Dosing Engine supporting fixed, variable, tapering, PRN, and range regimens with automated quantity calculation.


3. Multi-Role Care-Line Support, Talk-Through AI Assistant & Call Diarization System (Meta MMS-1B + Muse Spark):

- Built the Stage-Scoped Care-Line Conversation Tunnel ('clinical', 'nurse', 'pharmacy', 'delivery') with open/pause/close lifecycle controls, seamless in-call Audio/Video/Chat modality switching, and least-privilege privacy isolation between doctors, nurses, pharmacists, and couriers.

- Built the Floating Talk-Through AI Support Assistant (AICopilotFloatingPill.tsx & DashboardAIAssistant.tsx) across all 9 facility dashboards (Care Homes, Clinics, Hospitals, Pharmacies, PPMVs, Labs, Radiology, and Customer Care / Phone Desk) plus phone-line assisted visit scheduling.

- Deployed Meta MMS-1B + Pyannote 3.1 on Modal GPU for 2-way doctor-patient call audio transcription and speaker diarization directly into the EHR.


LIVE JUDGE ACCESS & DEMO CREDENTIALS

Live Platform: https://docarion.com

Login Page: https://docarion.com/auth/login

Dedicated Judge Login: judge@docarion.clinic (Password: Password123!)

Pre-Seeded Judge Consultation Room: https://docarion.com/consultations/consult-meta-judge-demo

Public Bilingual Symptom Router (No Login Required): https://docarion.com/symptom-checker

Additional Role Accounts (Password: Password123!):

- Doctor: doc1@docarion.clinic (/dashboard/doctor)

- Walk-In Nurse Care Station: nurse1@docarion.clinic (/care-station)

- PPMV Counter Attendant: attendant1@docarion.ppmv (/dashboard/pharmacy)

- Pharmacist: pharm1@docarion.pharmacy (/dashboard/pharmacy)

- Patient: patient1@docarion.clinic (/dashboard/patient)

## More from Meta Global AI Developer Hackathon

- [Royal Bengal](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/465)
- [MotionFlow Studio](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/466)
- [SalesCrew](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/467)
- [geometry os](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/469)
- [solo q](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/470)
- [RescueLens](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/471)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/468. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt