msw-auto-mock

A cli tool to generate random mock data from OpenAPI descriptions for msw.

Why

We already have all the type definitions from OpenAPI spec so hand-writing every response resolver is completely unnecessary.

Generative AI Support

Since v0.19.0, msw-auto-mock can use generative AI to generate mock data instead of faker.

AI is configured via your package.json (or any config file supported by cosmiconfig) under the msw-auto-mock key:

{
  "msw-auto-mock": {
    "ai": {
      "enable": true,
      "provider": "openai",
      "openai": {
        "apiKey": "process.env.OPENAI_API_KEY",
        "model": "gpt-4o"
      }
    }
  }
}

Providers and Required Fields

Currently supported providers: openai, azure, anthropic.

When ai.enable is true, you must provide the model/deployment for the selected provider:

• provider: "openai" requires ai.openai.model
• provider: "azure" requires ai.azure.deployment
• provider: "anthropic" requires ai.anthropic.model

If a required field is missing, generation fails fast with a clear error (instead of generating broken code).

How Config Values Are Interpreted

The AI config values are strings in JSON, but they are used to generate JavaScript code. msw-auto-mock supports two styles:

• Expression strings: values that start with process.env. or import.meta.env. (or Deno.env.) are treated as JavaScript expressions and injected as-is.
• Literal strings: everything else is treated as a string literal and will be automatically quoted in the generated code.

This means you can write:

• "model": "gpt-4o" (literal string; recommended)
• "apiKey": "process.env.OPENAI_API_KEY" (expression string)

Full Type Definition

interface Config {
  ai?: {
    enable?: boolean;
    provider: 'openai' | 'azure' | 'anthropic';
    openai?: {
      baseURL?: string;
      apiKey?: string;
      model?: string;
    };
    azure?: {
      apiKey?: string;
      resource?: string;
      deployment?: string;
    };
    anthropic?: {
      apiKey?: string;
      model?: string;
    };
  };
}

Important

For security, keep API keys in environment variables and reference them via process.env.* / import.meta.env.* in your config. You no longer need to wrap model names in extra quotes; "model": "gpt-4o" is valid.

Usage

This tool also requires @faker-js/faker >= 8 and msw >= 2.

Install:

yarn add msw-auto-mock @faker-js/faker -D

Read from your OpenAPI descriptions and output generated code:

# can be http url or a file path on your machine, support both yaml and json.
npx msw-auto-mock http://your_openapi.json -o ./mock

Integrate with msw, see Mock Service Worker's doc for more detail:

# Install msw
yarn add msw --dev

# Init service worker
npx msw init public/ --save

Then import those mock definitions in you app entry:

import { worker } from './mock/browser.js';

await worker.start();

For conditional mocking:

async function enableMocking() {
  if (process.env.NODE_ENV !== 'development') {
    return;
  }
  const { worker } = await import('./mock/browser');
  // `worker.start()` returns a Promise that resolves
  // once the Service Worker is up and ready to intercept requests.
  return worker.start();
}

function mountApp() {
  const root = createRoot(document.getElementById('root'));
  root.render(<App />);
}

enableMocking().then(mountApp);

For Node.js integration, you can import from your_output/node.js:

import { worker } from './mock/node.js';

For React Native integration, you can import from your_output/native.js:

import { worker } from './mock/native.js';

Run you app then you'll see a successful activation message from Mock Service Worker in your browser's console.

Options

• -o, --output: specify output file path or output to stdout.
• -m, --max-array-length <number>: specify max array length in response, default value is 20, it'll cost some time if you want to generate a huge chunk of random data.
• -t, --includes <keywords>: specify keywords to match if you want to generate mock data only for certain requests, multiple keywords can be seperated with comma.
• -e, --excludes <keywords>: specify keywords to exclude, multiple keywords can be seperated with comma.
• --base-url: output code with specified base url or fallback to server host specified in OpenAPI.
• --static: By default it will generate dynamic mocks, use this flag if you need it to be static.
• -c, --codes <keywords>: comma separated list of status codes to generate responses for
• --typescript: Generate TypeScript files instead of JavaScript files.
• -h, --help: show help info.