Patient-facing (coming soon)

Patient communication

Model: aug-patient-explain · Status: Coming soon · Scope: agents:patient_comms · Schema: document/0.1

Turns clinical content into clear, accurate patient communication — explanations of a condition or test result, medication instructions, procedure preparation and aftercare, discharge advice, and follow-up / adherence messages — in English and Indian languages (Hindi, Bengali, Tamil, Telugu, Marathi, Kannada, Malayalam, Gujarati, Punjabi, Odia).

Request

JSON

{
  "model": "aug-patient-explain",
  "messages": [{"role": "user", "content": "Explain the discharge plan"}],
  "augsidius": {
    "audience": "patient",
    "locale": "ta-IN",
    "context": {"conditions": [{"text": "community-acquired pneumonia"}], "medications": [{"text": "amoxicillin-clavulanate 625 mg TDS x 5 days"}]},
    "instructions": {"kind": "discharge_advice", "reading_level": "simple", "channel": "whatsapp", "max_chars": 900}
  }
}

kind: condition_explainer, result_explainer, medication_instructions, procedure_prep, procedure_aftercare, discharge_advice, follow_up_reminder, adherence_nudge.

Response payload

JSON

{
  "document": {"type": "discharge_advice", "language": "ta",
               "sections": [{"title": "…", "text": "…"}, {"title": "உடனே மருத்துவமனைக்கு வரவும்:", "items": ["…"]}]},
  "plain_text": "…",
  "reading_level": "grade_6"
}

Medical facts come from the clinical input and the evidence graph; the model rewrites, it does not add new treatment advice.
Doses are copied exactly from the prescription, never calculated.
Every message ends with when to seek help.

Triage

EMR integration