​
Overview

A PromptType consists of the following main components:

​
config
ConfigType

Provider and runtime configuration.

Constraints: Required configuration object.

​
messages
Array<MessageType>

Array of conversation messages.

Constraints: Required, at least one message.

​
tools
Array<ToolType>

Available tools for the LLM.

Constraints: Required array (can be empty)

​
variables
Array<VariableType>

Dynamic content variables.

Constraints: Optional, unique names required.

​
Field Details

​
config

For complete configuration details, see the Config Reference

​
messages

For complete message structure details, see the MessageType Reference

​
tools

For complete tool structure details, see the ToolType Reference

​
variables

Optional array of dynamic variables that can be substituted into prompt content.

​
name
string

Variable identifier.

Constraints: 1-200 characters, must be unique within prompt.

​
value
ContentType

The variable’s content.

Constraints: Any valid content type (text, image, etc.)

​
description
string

Optional variable description.

Constraints: Optional, 1-200 characters.

For complete ContentType details, see the ContentType Reference

​
Name Constraints

  • Length: 1-200 characters
  • Uniqueness: Must be unique within the same prompt
  • Usage: Can be referenced in message content for dynamic substitution

​
Complete Examples

​
Basic Prompt

const prompt = {
  config: {
    providerName: "openai",
    providerId: "6e94350a-95a9-48e7-8eab-a9f35ec2dc9d",
    model: "gpt-4o",
    settings: {
      temperature: 0.7,
      maxTokens: 500,
    },
  },
  messages: [
    {
      role: "system",
      content: [
        {
          modality: "text",
          value: "You are a helpful assistant.",
        },
      ],
    },
    {
      role: "user",
      content: [
        {
          modality: "text",
          value: "What is the weather like?",
        },
      ],
    },
  ],
  tools: [],
  variables: [],
};

​
Prompt with Tools and Variables

const prompt = {
  config: {
    providerName: "anthropic",
    providerId: "6e94350a-95a9-48e7-8eab-a9f35ec2dc9d",
    model: "claude-3-sonnet-20240229",
    settings: {
      temperature: 0.3,
      maxTokens: 2000,
    },
  },
  messages: [
    {
      role: "system",
      content: [
        {
          modality: "text",
          value:
            "You are an AI assistant that helps users with weather information. Always use the get_weather tool to provide accurate data.",
        },
      ],
    },
    {
      role: "user",
      content: [
        {
          modality: "text",
          value: "What's the weather like in {{location}}? I'm planning activities for {{activity_type}}.",
        },
      ],
    },
  ],
  tools: [
    {
      type: "function",
      definition: {
        schema: {
          name: "get_weather",
          description: "Get current weather information for a specific location",
          parameters: {
            type: "object",
            properties: {
              location: {
                type: "string",
                description: "City and state or city and country",
              },
              units: {
                type: "string",
                enum: ["celsius", "fahrenheit"],
                default: "fahrenheit",
              },
            },
            required: ["location"],
          },
        },
      },
    },
  ],
  variables: [
    {
      name: "location",
      value: {
        modality: "text",
        value: "San Francisco, CA",
      },
      description: "The location to get weather for",
    },
    {
      name: "activity_type",
      value: {
        modality: "text",
        value: "outdoor activities",
      },
      description: "Type of activities the user is planning",
    },
  ],
};

​
Prompt with Image Variable

const prompt = {
  config: {
    providerName: "openai",
    providerId: "6e94350a-95a9-48e7-8eab-a9f35ec2dc9d",
    model: "gpt-4o",
    settings: {
      temperature: 0.5,
      maxTokens: 1000,
    },
  },
  messages: [
    {
      role: "system",
      content: [
        {
          modality: "text",
          value: "You are an expert at analyzing images. Describe what you see in detail.",
        },
      ],
    },
    {
      role: "user",
      content: [
        {
          modality: "text",
          value: "Please analyze this image and tell me about {{subject}}:",
        },
        {
          modality: "image",
          detail: "high",
          value: {
            type: "url",
            url: "{{image_url}}",
          },
        },
      ],
    },
  ],
  tools: [],
  variables: [
    {
      name: "subject",
      value: {
        modality: "text",
        value: "the architectural style",
      },
      description: "What aspect to focus on in the analysis",
    },
    {
      name: "image_url",
      value: {
        modality: "text",
        value: "https://example.com/building.jpg",
      },
      description: "URL of the image to analyze",
    },
  ],
};

​
Multi-Variable Prompt

const prompt = {
  config: {
    providerName: "google",
    providerId: "6e94350a-95a9-48e7-8eab-a9f35ec2dc9d",
    model: "gemini-2.5-flash-exp",
    settings: {
      temperature: 0.4,
      maxTokens: 1500,
    },
  },
  messages: [
    {
      role: "system",
      content: [
        {
          modality: "text",
          value: "You are {{assistant_role}}. Always be {{personality_trait}} and provide {{response_style}} responses.",
        },
      ],
    },
    {
      role: "user",
      content: [
        {
          modality: "text",
          value: "Hello! I'm {{user_name}} and I need help with {{task_description}}. I have {{time_constraint}} to complete this.",
        },
      ],
    },
  ],
  tools: [
    {
      type: "function",
      definition: {
        schema: {
          name: "schedule_task",
          description: "Schedule a task with deadline",
          parameters: {
            type: "object",
            properties: {
              task: { type: "string" },
              deadline: { type: "string" },
              priority: {
                type: "string",
                enum: ["low", "medium", "high"],
              },
            },
            required: ["task", "deadline"],
          },
        },
      },
    },
  ],
  variables: [
    {
      name: "assistant_role",
      value: {
        modality: "text",
        value: "a productivity expert",
      },
      description: "The role the assistant should take",
    },
    {
      name: "personality_trait",
      value: {
        modality: "text",
        value: "encouraging and supportive",
      },
      description: "The personality the assistant should exhibit",
    },
    {
      name: "response_style",
      value: {
        modality: "text",
        value: "actionable and concise",
      },
      description: "The style of responses to provide",
    },
    {
      name: "user_name",
      value: {
        modality: "text",
        value: "Alex",
      },
      description: "The user's name",
    },
    {
      name: "task_description",
      value: {
        modality: "text",
        value: "organizing my project timeline",
      },
      description: "Description of what the user needs help with",
    },
    {
      name: "time_constraint",
      value: {
        modality: "text",
        value: "3 days",
      },
      description: "How much time the user has",
    },
  ],
};