DialogMD Documentation

Welcome to DialogMD, a HIPAA-compliant platform for building intelligent healthcare chatflows without code.

Key Features

• Visual Flow Builder: Drag-and-drop interface for creating complex conversation flows
• HIPAA Compliant: Built-in security features and audit logging
• AI-Powered Agents: Integrate with GPT-4, Claude, and other LLMs
• Custom Integrations: Connect to your EHR, scheduling system, and APIs
• Multi-Tenant Architecture: Manage multiple organizations with data isolation

Getting Started

Follow these steps to create your first chatflow in DialogMD.

Flow Builder

The Flow Builder is the heart of DialogMD. Use it to visually design conversation flows by connecting nodes together.

Interface Overview

• Canvas: The main work area where you place and connect nodes
• Node Palette: Left sidebar containing all available node types
• Properties Panel: Configure selected node settings
• Test Panel: Right sidebar for testing your flow in real-time
• Toolbar: Save, publish, and manage flow settings

Creating a Flow

1. Drag nodes: Drag node types from the palette onto the canvas
2. Connect nodes: Click and drag from a node's output handle to another node's input
3. Configure nodes: Double-click a node or click the edit icon to open properties
4. Test flow: Use the test panel to simulate conversations
5. Save and publish: Save your work and publish when ready

Node Types

Message Node

Displays a message to the user. Supports markdown formatting, variable interpolation, and media.

Example Use Cases

• Welcome messages
• Instructions and explanations
• Confirmation messages
• Dynamic content with variable data

User Input Node

Collects input from users with support for multiple input types including text, email, phone, date, ratings, and more.

Agent Node

Integrates AI agents powered by LLMs like GPT-4, Claude, or Llama. Configure system prompts, temperature, and tools for intelligent responses.

Use Cases

• Symptom triage and assessment
• Answering patient questions
• Data extraction and classification
• Appointment scheduling assistance

Function Node

Execute custom JavaScript code to manipulate variables, perform calculations, or implement custom logic.

// Calculate age from birthdate
const birthdate = getVariable('birthdate')
const today = new Date()
const birth = new Date(birthdate)
const age = today.getFullYear() - birth.getFullYear()
setVariable('age', age)

// Determine if patient is eligible
if (age >= 18) {
  setVariable('eligible', true)
} else {
  setVariable('eligible', false)
}

API Node

Make HTTP requests to external APIs or your EHR system. Extract data from responses and store in variables.

Branching Node

Create conditional logic to route conversations based on variable values or conditions.

// Example conditions:
variables.age >= 18
variables.symptoms.includes('fever')
variables.urgency === 'high'

Card Carousel Node

Display a carousel of cards for users to select from. Perfect for showing options, services, or products.

Appointment Booking Node

Enable patients to book appointments directly in the conversation. Integrates with your scheduling API.

Agents

Agents are AI-powered assistants that can understand natural language and respond intelligently. Configure agents in your library and reuse them across multiple flows.

Creating an Agent

1. Navigate to Agents: Go to Dashboard → Agents
2. Click "Create Agent"
3. Configure Agent:
   • Name and description
   • Select LLM model (GPT-4, Claude, etc.)
   • Set system prompt
   • Configure temperature and max tokens
   • Add tools and functions if needed
4. Test: Use the test interface to validate responses
5. Save: Agent is now available in your library

Best Practices

• Clear Instructions: Write detailed system prompts that explain the agent's role
• Context: Provide relevant context about the conversation in the agent node
• Temperature: Use lower values (0.1-0.3) for factual responses, higher (0.7-0.9) for creative tasks
• Safety: Include guidelines about HIPAA compliance and PHI handling in prompts

Environments

Environments allow you to maintain separate configurations for development, staging, and production.

Environment Types

Credentials

Securely store API keys, tokens, and authentication credentials for your integrations.

Credential Types

• API Key: Simple key-based authentication
• Bearer Token: OAuth bearer tokens
• Basic Auth: Username and password
• OAuth 2.0: Full OAuth flow credentials
• Custom: Custom header-based authentication

API Configurations

Configure external APIs to integrate with your EHR, scheduling system, or other healthcare services.

Setting Up an API

1. Create Configuration: Name your API and provide base URL
2. Add Credentials: Link authentication credentials
3. Configure Endpoints: Define endpoints, methods, and parameters
4. Test Connection: Verify API connectivity
5. Environment Setup: Configure for each environment

MCP Servers & Tools

Model Context Protocol (MCP) servers provide tools and resources that AI agents can use to access external data and perform actions.

What is MCP?

MCP is a standardized protocol for connecting AI models to external tools, databases, and APIs. It enables agents to perform actions like searching databases, calling APIs, or accessing files.

Use Cases

• Give agents access to patient databases
• Enable appointment booking capabilities
• Connect to medical knowledge bases
• Integrate with lab result systems

Themes

Customize the look and feel of your chatflows to match your brand.

Theme Options

Security & Compliance

DialogMD is built with healthcare security and HIPAA compliance as top priorities.

HIPAA Compliance

• End-to-end encryption for all data in transit and at rest
• Comprehensive audit logging of all PHI access
• Role-based access control (RBAC)
• Automatic session timeout and re-authentication
• Business Associate Agreement (BAA) included

Data Protection

• Multi-tenant Isolation: Complete data separation between organizations
• Encryption: AES-256 encryption for data at rest, TLS 1.3 for data in transit
• Backup & Recovery: Automated daily backups with point-in-time recovery
• Monitoring: 24/7 security monitoring and threat detection

Best Practices

Flow Design

• Keep it Simple: Start with the happy path, then add edge cases
• Clear Messages: Use plain language and avoid medical jargon when possible
• Error Handling: Always include error nodes and provide helpful messages
• Test Thoroughly: Test all conversation paths before publishing
• User Experience: Don't ask for information you already have

Variables

• Naming: Use descriptive camelCase names (e.g., patientFirstName)
• Scope: Variables persist throughout the conversation
• Privacy: Be mindful of PHI in variable names and values
• Validation: Always validate user input before storing

Performance

• API Calls: Minimize unnecessary API calls
• Agent Usage: Use agents strategically for complex tasks only
• Caching: Cache static data when possible
• Testing: Use development environment for testing