Seeded intake journeys
Select a patient scenario, then click the chat to walk through the flow. All four panels update live.
Patient intake chat
Claude-style guided questions · PHI redacted on egressPick a seeded patient scenario above and press "Next message" to walk through an intake. Every turn emits audit entries on the right.
Structured capture
JSON fields populate live as the LLM extracts them0 / 10 fields{
"chief_complaint": null,
"symptoms[]": null,
"onset": null,
"duration_days": null,
"severity": null,
"red_flags[]": null,
"medications[]": null,
"allergies[]": null,
"prior_conditions[]": null,
"patient_age_range": null
}
Triage routing · decision tree
highlighted path shows this patient's traversaldepth 0
Intake received
depth 1
Red flags detected
No red flags
depth 2
Cardiac / respiratory red flag
Other red flag
Symptoms > 14 days
Symptoms 3-14 days
Symptoms < 3 days
depth 3
Route → Urgent Care
Mechanical joint pattern
Dermatologic pattern
Moderate severity
Mild severity
Route → PCP
depth 4
Route → Orthopedics
Route → Dermatology
Appointment booking
awaiting routing decisionComplete the intake to see matching slots.
Mock EHR · encrypted writes
padlock = per-field envelope encryptionNo EHR records yet. A booked appointment writes an encrypted SOAP note here.