Package Exports

mindstudio
mindstudio/dist/index.js
mindstudio/package.json

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (mindstudio) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

MindStudio

Client library for MindStudio AI Workers. Easily integrate and execute AI workflows in your applications with type-safe interfaces.

Installation

npm install mindstudio

Quick Start

import { MindStudio } from 'mindstudio';

// Initialize the client
const client = new MindStudio('your-api-key');
await client.init();

// Execute a workflow
const { success, result } = await client.workers.myWorker.generateText({
  prompt: "Write a story about a space cat"
});

if (success) {
  console.log(result);
}

Configuration

Create a .env file in your project root:

MINDSTUDIO_KEY=your-api-key
MINDSTUDIO_BASE_URL=https://custom-api-endpoint.com  # Optional

Or pass configuration when initializing:

const client = new MindStudio('your-api-key', {
  baseUrl: 'https://custom-api-endpoint.com'
});

Type Definitions

The library provides comprehensive TypeScript definitions. After initialization, you'll get full type support for your workflows:

// Import available types
import { 
  WorkflowResponse, 
  OutputVarsResponse, 
  StringResponse,
  MindStudioWorkers 
} from 'mindstudio';

// Type-safe worker instance
const client: MindStudio;
const workers: MindStudioWorkers;

// Workflow with output variables
const response: OutputVarsResponse<{
  blogPost: string;
  title: string;
}> = await client.workers.contentGenerator.createBlogPost({
  topic: "AI Technology",
  tone: "Professional"
});

// Workflow with string output
const response: StringResponse = await client.workers.textGenerator.generateText({
  prompt: "Write a story"
});

Response Types

All workflow executions return a consistent structure:

// Base response type
interface WorkflowResponse<TResult> {
  success: boolean;
  result: TResult;
  error?: any;
  billingCost?: number;
}

// For workflows with output variables
type OutputVarsResponse<T extends Record<string, string>> = WorkflowResponse<T>;

// For workflows with string output
type StringResponse = WorkflowResponse<string | undefined>;

CLI Usage

MindStudio includes a CLI for workspace management:

Initialize Workspace

npx mindstudio sync

This will:

Create .mindstudio.json with your worker configurations
Generate TypeScript definitions in node_modules/mindstudio/dist/generated.d.ts
Set up your development environment

Test Workflows

npx mindstudio test [--worker <name>] [--workflow <name>] [--input <json>]

Interactive CLI to test workflows with input validation.

Regenerate Types

npx mindstudio generate

Updates TypeScript definitions from your current .mindstudio.json configuration.

Error Handling

import { MindStudioError } from 'mindstudio';

try {
  const { success, result, error } = await client.workers.myWorker.generateText({
    prompt: "Hello"
  });

  if (!success) {
    console.error('Workflow failed:', error);
    return;
  }
} catch (error) {
  if (error instanceof MindStudioError) {
    console.error({
      code: error.code,      // Error code string
      status: error.status,  // HTTP status code
      message: error.message,// Error description
      details: error.details // Additional context
    });
  }
}

Best Practices

Version Control

Commit .mindstudio.json to version control to ensure:

Consistent worker configurations across your team
Type definitions can be regenerated on new installations
AI workflow changes can be tracked

Post-Install Type Generation

Add to your package.json:

{
  "scripts": {
    "postinstall": "mindstudio generate"
  }
}

This regenerates types automatically when:

Installing on a new machine
Running npm install
Updating the mindstudio package

Note: Requires .mindstudio.json in your repository.

License

MIT