Best viewed on a computer

This interactive demo is built for a larger screen. For the full experience, open it on a laptop or desktop.

primeVCTR / LabSync — Data Architecture

From a lab PDF
to a pre-visit summary in three acts.

A live trace of how LabSync turns an outside-lab scan into a clinician-ready FHIR record. LabSync ingests and orchestrates. AWS extracts, stores, reasons, and serves.

01
Steps 01–05
Upload
Textract OCR → LabSync structures → writes Patient, Observations, DocumentReference. Persona · MA
02
Steps 06–07
Insights
Patient Insights auto-fires. PRE_VISIT narrative + sparklines, all evidenced. Server-driven
03
Steps 08–10
Review
Doctor opens /clinician. Three clicks from sentence to source PDF. Persona · Doctor
Users LabSync platform EHR · vendor sidecar AWS · us-east-1 upload · 2s poll page load Textract OCR roster + schedule PDF · cached PI FHIR PUT/GET StartPatientInsightsJob summary.json PI reads patient bundle from HealthLake presigned URL · source PDF
Persona
MA / Front Desk
Drops PDFs at /queue
Watches: parsing → ready
Persona
Doctor
Reviews /clinician
Clicks evidence chips
PrimeVCTR LabSync
App Server
PDF extraction · AWS orchestration
SSE publisher · evidence aggregator
Vendor sidecar
EHR (mock)
Schedule · roster · encounters
Epic / eCW / Athena / Cerner
AWS HealthLake
FHIR Datastore
Patient · Observation[]
DocumentReference
SMART-on-FHIR · Cognito JWT
System of record
AWS Textract
OCR
analyze_document
Text + tables from PDF
First AWS call
AWS S3
Object Storage
Source PDFs — Quest / LabCorp scans
PI outputsummary.json
Presigned URLs for browser reads
AWS Connect Health
Patient Insights
StartPatientInsightsJob · PRE_VISIT
Reads HealthLake bundle
Writes summary.json to S3
Auto-fires on upload
01/ 10
Act 01 Persona · MA Drop PDF
A lab PDF is dropped in /queue.
Browser POSTs the PDF to /api/upload. The LabSync server receives it. LabSync only — no AWS yet.
Deterministic ids · idempotent replays · HealthLake = system of record · S3 = source + derived artifacts
01 / 10
Persona · MA A lab PDF is dropped in /queue.
Browser POSTs the PDF to /api/documents/upload. The LabSync server receives it. LabSync only — no AWS yet.