Tool Types

This page documents the interfaces and types used for App tools integration with Mira AI.

For a complete guide on implementing App tools, see AI Tools.

ToolSchema

Interface defining the structure of a tool that a App can expose to Mira AI.

interface ToolSchema {
  /** Unique identifier for the tool */
  id: string

  /** Human-readable description of what the tool does */
  description: string

  /** Optional phrases that might trigger this tool (helps Mira recognize when to use it) */
  activationPhrases?: string[]

  /** Definition of parameters this tool accepts */
  parameters?: Record<string, ToolParameterSchema>
}

ToolParameterSchema

Interface defining the structure of parameters that a tool accepts.

interface ToolParameterSchema {
  /** Data type of the parameter */
  type: "string" | "number" | "boolean"

  /** Human-readable description of what the parameter is for */
  description: string

  /** Optional list of allowed values for string parameters */
  enum?: string[]

  /** Whether this parameter is required */
  required?: boolean
}

ToolCall

Interface representing a call to a App tool from Mira AI.

interface ToolCall {
  /** ID of the tool being called */
  toolId: string

  /** Parameter values for this specific call */
  toolParameters: Record<string, string | number | boolean>

  /** When the tool call was made */
  timestamp: Date

  /** ID of the user who triggered the tool call */
  userId: string

  /** The active session for the user who triggered the tool call, if the user is currently running this app */
  activeSession: AppSession | null
}

GIVE_APP_CONTROL_OF_TOOL_RESPONSE

The string GIVE_APP_CONTROL_OF_TOOL_RESPONSE is a special string that can be returned by your app to indicate that Mira should not respond to the user, and your app will respond directly.

import {GIVE_APP_CONTROL_OF_TOOL_RESPONSE} from "@mentra/sdk"

export class TodoAppServer extends AppServer {
  protected async onToolCall(toolCall: ToolCall): Promise<string | undefined> {
    const {toolId, activeSession} = toolCall

    // Handle different tool calls
    switch (toolId) {
      case "get_todos": {
        const todoList = userTodos
          .map(todo => `- ${todo.completed ? "✓" : "○"} ${todo.text}${todo.dueDate ? ` (due ${todo.dueDate})` : ""}`)
          .join("\n")

        if (activeSession) {
          // if the user is currently using the app, display the todo list in the app directly
          activeSession.layouts.showTextWall(todoList)
          return GIVE_APP_CONTROL_OF_TOOL_RESPONSE
        } else {
          // if the user is not currently using the app, return the todo list to Mira for Mira to relay
          return `Your todo list:\n${todoList}`
        }
      }
    }
  }
}

Tool Configuration

Tools are defined in the devloper console. Go to console.mentra.glass/apps and edit your App, then look for the "AI Tools" section.

Each tool definition has:

• id: Unique identifier for the tool
• description: Human/AI-readable description of what the tool does
• activationPhrases: Optional comma-separated list of phrases that might trigger this tool (although Mira may also trigger tools based on the context of the conversation)
• parameters: Optional list of parameters the tool accepts

Parameter Properties

Each parameter definition has:

• type: Data type of the parameter - "string", "number", or "boolean"
• description: Human/AI-readable description of the parameter
• required: Whether the parameter is required
• enum: Optional comma-separated list of allowed values for string parameters (if specified, Mira will choose one of these values)