Prompt Guide

Write better prompts, get better healthcare responses - patterns and defaults for every Vaidya use case.

What this page is for

This guide helps you write effective prompts for Vaidya-v2 to get accurate, safe, and actionable healthcare responses. You'll learn how to structure requests with clear objectives, patient context, input data, and output requirements.

Rules of thumb

Be specific - Provide specific instructions and constraints in the prompt to guide the model's response. For example, specifying the desired format (e.g., bullet points, word limit) can help ensure the output is structured in a way that is most useful for your application.
Control temperature - For factual medical tasks, use 0.2–0.3 for more deterministic responses. Higher values can be used for brainstorming or personalized plans, but always test and validate outputs carefully.
Add Context - Include relevant patient information wherever necessary for the best outputs.

Example Use Cases and Prompts

1. Symptom Checker

Task:
Triage symptoms and provide safe next steps.

User Profile:
- Age/Sex: <value>
- Known Conditions: <value>
- Current Medications: <value>
- Allergies: <value>

Context:
- Onset: <value>
- Duration/trend: <value>
- Severity: <mild/moderate/severe>
- Symptoms: <list>
- Vitals: <if available>

Suggest the next best step: Home Remedies / PCP visit / Urgent care / ER
Also give a brief rationale for the recommendation.

2. Patient Journey – Personalized Health Plan

Task:
Build a practical personalized plan for weight loss

# User Profile:
- Age/Sex: <value>
- Conditions: <value>
- Work/schedule constraints: <value>
- Diet preference: <value>

# Output Instructions
1. Output weekly exercise plan balancing cardio and strength, with 3-5 specific activities per week.
2. Provide daily meal suggestions with portion sizes, focusing on whole foods and balanced macros. Only give vegetarian options.

Multi-turn chats

Keep clinically relevant turns in messages.
Preserve role order.
Long thread? Summarize into one compact user message for better context retention.

Example: Multi-turn Input for Postpartum Recovery Assessment

{
  "model": "Vaidya-v2",
  "messages": [
    {
      "role": "user",
      "content": "Task: Assess postpartum recovery. User Profile: 32/F, 2 weeks postpartum, vaginal delivery. Heavy vaginal bleeding, mild lower abdominal cramping, no fever"
    },
    {
      "role": "assistant",
      "content": "Heavy bleeding at 2 weeks is concerning; normal at this stage is moderate bleeding with small clots. Mild cramping is expected. Absence of fever is reassuring (no infection signs). Red flags: soaking >1 pad/hour, foul-smelling discharge, severe pain, fever >101°F, dizziness. Recommend: immediate postpartum follow-up, assess hemoglobin if not done, rule out retained tissue."
    },
    {
      "role": "user",
      "content": "She's feeling dizzy when standing up. No fever. Last postpartum checkup was at 1 week and everything was fine then. Hemoglobin was 10.2 at discharge (down from 12.5 pre-pregnancy)."
    }
  ]
}

Standard text request

{
  "model": "Vaidya-v2",
  "messages": [
    {
      "role": "user",
      "content": "My neighbour has severe chest pain and is feeling dizzy. What should I do?"
    }
  ],
  "stream": false
}