Zimic

TypeScript-first HTTP request mocking

npm   •   Docs   •   Issues   •   Roadmap

🚧 This project is under active development! Check our roadmap to v1. Contributors and ideas are welcome!

Zimic is a lightweight, TypeScript-first HTTP request mocking library, inspired by Zod's type inference and using MSW under the hood.

Features

Zimic provides a simple, flexible and type-safe way to mock HTTP requests.

  • Statically-typed mocks. Declare your HTTP endpoints and get full type inference and validation when applying mocks.
  • 🔗 Network-level intercepts. Internally, Zimic uses MSW, which intercepts HTTP requests right before they leave your app. This means that no parts of your code are stubbed or skipped. From you application's point of view, the mocked requests are indistinguishable from the real ones.
  • 🔧 Flexibility. You can simulate real application workflows by mocking each necessary endpoints. This is specially useful in testing, making sure the real path your application takes is covered and the sent requests are correct.
  • 💡 Simplicity. Zimic was designed from scratch to encourage clarity, simplicity and standardization in your mocks. Check our Getting started guide and starting mocking!

Table of contents

Getting started

1. Requirements

  • TypeScript >=4.7

  • strict mode enabled in your tsconfig.json:

    {
      // ...
      "compilerOptions": {
        // ...
        "strict": true,
      },
    }

2. Install from npm

Manager Command
npm npm install zimic --save-dev
yarn yarn add zimic --dev
pnpm pnpm add zimic --dev
bun bun add zimic --dev

Canary versions are released under the canary tag:

Manager Command
npm npm install zimic@canary --save-dev
yarn yarn add zimic@canary --dev
pnpm pnpm add zimic@canary --dev
bun bun add zimic@canary --dev

3. Post-install

Node.js post-install

No additional configuration is necessary for Node.js. Check out the usage guide and start mocking!

Browser post-install

To use Zimic on a browser, you should first initialize the mock service worker in your public directory:

npx zimic browser init <publicDirectory>

This will create a mockServiceWorker.js file in the provided public directory, which is necessary to intercept requests and mock responses. We recommend not deploying this file to production.

Usage

Examples

Visit our examples to see how to use Zimic with popular frameworks and libraries!

Basic usage

To start using Zimic, create a worker targeting your platform.

import { createHttpInterceptorWorker } from 'zimic/interceptor';

const worker = createHttpInterceptorWorker({
  platform: 'node', // or 'browser'
});

Then, create your first interceptor:

import { createHttpInterceptor } from 'zimic/interceptor';

const interceptor = createHttpInterceptor<{
  '/users': {
    GET: {
      response: {
        200: { body: User[] };
      };
    };
  };
}>({
  worker,
  baseURL: 'http://localhost:3000',
});

In this example, we're creating an interceptor for a service with a single path, /users, that supports a GET method. The response for a successful request is an array of User objects. Learn more about declaring interceptor schemas at HttpInterceptor schema.

Finally, start the worker to intercept requests:

await worker.start();

Now, you can start intercepting requests and returning mock responses!

const listTracker = interceptor.get('/users').respond({
  status: 200,
  body: [{ username: 'diego-aquino' }],
});

const response = await fetch('http://localhost:3000/users');
const users = await response.json();
console.log(users); // [{ username: 'diego-aquino' }]

More examples are available at zimic/interceptor API.

Testing

We recommend managing the lifecycle of your workers and interceptors using beforeAll and afterAll hooks in your test setup file. An example using Jest/Vitest structure:

// tests/setup.ts
import { createHttpInterceptorWorker, createHttpInterceptor } from 'zimic/interceptor';

// create a worker
const worker = createHttpInterceptorWorker({
  platform: 'node',
});

const userInterceptor = createHttpInterceptor<{
  // declare your schema
}>({
  worker,
  baseURL: 'http://localhost:3000',
});

const logInterceptor = createHttpInterceptor<{
  // declare your schema
}>({
  worker,
  baseURL: 'http://localhost:3001',
});

beforeAll(async () => {
  // start intercepting requests
  await worker.start();
});

beforeEach(async () => {
  // clear all interceptors to make sure no tests affect each other
  userInterceptor.clear();
  logInterceptor.clear();
});

afterAll(async () => {
  // stop intercepting requests
  await worker.stop();
});

zimic API

This module provides general utilities, such as HTTP classes.

HttpHeaders

A superset of the built-in Headers class, with a strictly-typed schema. HttpHeaders is fully compatible with Headers and is used by Zimic abstractions to provide type safety when applying mocks.

import { HttpHeaders } from 'zimic';

const headers = new HttpHeaders<{
  accept?: string;
  'content-type'?: string;
}>({
  accept: '*/*',
  'content-type': 'application/json',
});

const contentType = headers.get('content-type');
console.log(contentType); // 'application/json'

HttpSearchParams

A superset of the built-in URLSearchParams class, with a strictly-typed schema. HttpSearchParams is fully compatible with URLSearchParams and is used by Zimic abstractions to provide type safety when applying mocks.

import { HttpSearchParams } from 'zimic';

const searchParams = new HttpSearchParams<{
  names?: string[];
  page?: `${number}`;
}>({
  names: ['user 1', 'user 2'],
  page: '1',
});

const names = searchParams.getAll('names');
console.log(names); // ['user 1', 'user 2']

const page = searchParams.get('page');
console.log(page); // '1'

zimic/interceptor API

This module provides a set of utilities to create HTTP interceptors for both Node.js and browser environments.

All APIs are documented using JSDoc comments, so you can view detailed descriptions directly in your IDE!

HttpInterceptorWorker

Workers are used by HttpInterceptor's to intercept HTTP requests and return mock responses. To start intercepting requests, the worker must be started.

In a project, all interceptors can share the same worker.

createHttpInterceptorWorker

Creates an HTTP interceptor worker. A platform is required to specify the environment where the worker will be running:

  • Node.js:

    import { createHttpInterceptorWorker } from 'zimic/interceptor';
    
    const worker = createHttpInterceptorWorker({
      platform: 'node',
    });
  • Browser:

    import { createHttpInterceptorWorker } from 'zimic/interceptor';
    
    const worker = createHttpInterceptorWorker({
      platform: 'browser',
    });

worker.platform()

Returns the platform used by the worker (browser or node).

const platform = worker.platform();

worker.start()

Starts the worker, allowing it to be used by interceptors.

await worker.start();

When targeting a browser environment, make sure to run npx zimic browser init <publicDirectory> on your terminal before starting the worker. This initializes the mock service worker in your public directory. Learn more at zimic browser init <publicDirectory.

worker.stop()

Stops the worker, preventing it from being used by interceptors.

await worker.stop();

worker.isRunning()

Returns whether the worker is currently running and ready to use.

const isRunning = worker.isRunning();

HttpInterceptor

HTTP interceptors provide the main API to handle HTTP requests and return mock responses. The methods, paths, status codes, parameters, and responses are statically-typed based on the service schema. To intercept HTTP requests, an interceptor needs a running HttpInterceptorWorker.

Each interceptor represents a service and can be used to mock its paths and methods.

createHttpInterceptor

Creates an HTTP interceptor, the main interface to intercept HTTP requests and return responses. Learn more about interceptor schemas at HttpInterceptor schema.

import { createHttpInterceptorWorker, createHttpInterceptor } from 'zimic/interceptor';

const worker = createHttpInterceptorWorker({
  platform: 'node',
});

const interceptor = createHttpInterceptor<{
  '/users/:id': {
    GET: {
      response: {
        200: { body: User };
        404: { body: NotFoundError };
      };
    };
  };
}>({
  worker,
  baseURL: 'http://localhost:3000',
});

HttpInterceptor schema

HTTP interceptor schemas define the structure of the real services being mocked. This includes paths, methods, request and response bodies, and status codes. Based on the schema, interceptors will provide type validation when applying mocks.

An example of a complete interceptor schema:
import { createHttpInterceptor } from 'zimic/interceptor';

const interceptor = createHttpInterceptor<{
  '/users': {
    POST: {
      request: {
        body: {
          username: string;
        };
      };
      response: {
        201: { body: User };
      };
    };
    GET: {
      response: {
        200: { body: User[] };
        404: { body: NotFoundError };
      };
    };
  };

  '/users/:id': {
    GET: {
      response: {
        200: { body: User };
        404: { body: NotFoundError };
      };
    };
  };

  '/posts': {
    GET: {
      response: {
        200: { body: Post[] };
      };
    };
  };
}>({
  worker,
  baseURL: 'http://localhost:3000',
});
Alternatively, you can compose the schema using utility types:
import { createHttpInterceptor, HttpInterceptorSchema } from 'zimic/interceptor';

type UserPaths = HttpInterceptorSchema.Root<{
  '/users': {
    POST: {
      request: {
        body: {
          username: string;
        };
      };
      response: {
        201: { body: User };
      };
    };

    GET: {
      response: {
        200: { body: User[] };
        404: { body: NotFoundError };
      };
    };
  };
}>;

type UserByIdPaths = HttpInterceptorSchema.Root<{
  '/users/:id': {
    GET: {
      response: {
        200: { body: User };
        404: { body: NotFoundError };
      };
    };
  };
}>;

type PostPaths = HttpInterceptorSchema.Root<{
  '/posts': {
    GET: {
      response: {
        200: { body: Post[] };
      };
    };
  };
}>;

type InterceptorSchema = HttpInterceptorSchema.Root<UserPaths & UserByIdPaths & PostPaths>;

const interceptor = createHttpInterceptor<InterceptorSchema>({
  worker,
  baseURL: 'http://localhost:3000',
});
Declaring paths

At the root level, each key represents a path or route:

import { createHttpInterceptor } from 'zimic/interceptor';

const interceptor = createHttpInterceptor<{
  '/users': {
    // path schema
  };
  '/users/:id': {
    // path schema
  };
  '/posts': {
    // path schema
  };
}>({
  worker,
  baseURL: 'http://localhost:3000',
});
Alternatively, you can also compose root level paths using the utility type HttpInterceptorSchema.Root:
import { createHttpInterceptor, HttpInterceptorSchema } from 'zimic/interceptor';

type UserPaths = HttpInterceptorSchema.Root<{
  '/users': {
    // path schema
  };
  '/users/:id': {
    // path schema
  };
}>;

type PostPaths = HttpInterceptorSchema.Root<{
  '/posts': {
    // path schema
  };
}>;

const interceptor = createHttpInterceptor<UserPaths & PostPaths>({
  worker,
  baseURL: 'http://localhost:3000',
});
Declaring methods

Each path can have one or more methods, (GET, POST, PUT, PATCH, DELETE, HEAD, and OPTIONS). The method names are case-sensitive.

import { createHttpInterceptor } from 'zimic/interceptor';

const interceptor = createHttpInterceptor<{
  '/users': {
    GET: {
      // method schema
    };
    POST: {
      // method schema
    };
  };
  // other paths
}>({
  worker,
  baseURL: 'http://localhost:3000',
});
Similarly to paths, you can also compose methods using the utility type HttpInterceptorSchema.Method:
import { createHttpInterceptor, HttpInterceptorSchema } from 'zimic/interceptor';

type UserMethods = HttpInterceptorSchema.Method<{
  GET: {
    // method schema
  };
  POST: {
    // method schema
  };
}>;

const interceptor = createHttpInterceptor<{
  '/users': UserMethods;
}>({
  worker,
  baseURL: 'http://localhost:3000',
});
Declaring requests

Each method can have a request, which defines the schema of the accepted requests. headers, searchParams, and body are supported to provide type safety when applying mocks.

Tip: You only need to declare the properties you want to use in your mocks. For example, if you are referencing only the header accept, you need to declare only it. Similarly, if you are not using any headers, search params, or body, you can omit them.

import { createHttpInterceptor } from 'zimic/interceptor';

const interceptor = createHttpInterceptor<{
  '/users': {
    POST: {
      request: {
        headers: {
          accept: string;
        };
        body: {
          username: string;
        };
      };
      // ...
    };

    GET: {
      request: {
        searchParams: {
          username?: string;
        };
      };
      // ...
    };

    // other methods
  };
  // other paths
}>({
  worker,
  baseURL: 'http://localhost:3000',
});
You can also compose requests using the utility type HttpInterceptorSchema.Request, similarly to methods:
import { createHttpInterceptor, HttpInterceptorSchema } from 'zimic/interceptor';

type UserCreationRequest = HttpInterceptorSchema.Request<{
  headers: {
    accept: '*/*';
  };
  body: {
    username: string;
  };
}>;

const interceptor = createHttpInterceptor<{
  '/users': {
    POST: {
      request: UserCreationRequest;
    };
  };
}>({
  worker,
  baseURL: 'http://localhost:3000',
});
Declaring responses

Each method can also have a response, which defines the schema of the returned responses. The status codes are used as keys. headers and body are supported to provide type safety when applying mocks.

Tip: Similarly to Declaring requests, you only need to declare the properties you want to use in your mocks. For example, if you are referencing only the header content-type, you can declare only it. If you are not using any headers or body, you can omit them.

import { createHttpInterceptor } from 'zimic/interceptor';

const interceptor = createHttpInterceptor<{
  '/users/:id': {
    GET: {
      // ...
      response: {
        200: {
          headers: {
            'content-type': string;
          };
          body: User;
        };
        404: {
          headers: {
            'content-type': string;
          };
          body: NotFoundError;
        };
      };
    };
    // other methods
  };
  // other paths
}>({
  worker,
  baseURL: 'http://localhost:3000',
});
You can also compose responses using the utility types HttpInterceptorSchema.ResponseByStatusCode and HttpInterceptorSchema.Response, similarly to requests:
import { createHttpInterceptor, HttpInterceptorSchema } from 'zimic/interceptor';


type SuccessUserGetResponse = HttpInterceptorSchema.Response<{
  headers: {
    'content-type': string;
  };
  body: User;
}>;

type NotFoundUserGetResponse = **HttpInterceptorSchema**.Response<{
  headers: {
    'content-type': string;
  };
  body: NotFoundError;
}>;

type UserGetResponses = HttpInterceptorSchema.ResponseByStatusCode<{
  200: SuccessUserGetResponse;
  404: NotFoundUserGetResponse;
}>;

const interceptor = createHttpInterceptor<{
  '/users/:id': {
    GET: {
      // ...
      response: UserGetResponses;
    };
    // other methods
  };
  // other paths
}>({
  worker,
  baseURL: 'http://localhost:3000',
});

interceptor.baseURL()

Returns the base URL of the interceptor.

const baseURL = interceptor.baseURL();

interceptor.<method>(path)

Creates an HttpRequestTracker for the given method and path. The path and method must be declared in the interceptor schema.

The supported methods are: get, post, put, patch, delete, head, and options.

import { createHttpInterceptor } from 'zimic/interceptor';

const interceptor = createHttpInterceptor<{
  '/users': {
    GET: {
      response: {
        200: { body: User[] };
      };
    };
  };
}>({
  worker,
  baseURL: 'http://localhost:3000',
});

// intercept any GET requests to http://localhost:3000/users and return this response
const listTracker = interceptor.get('/users').respond({
  status: 200
  body: [{ username: 'diego-aquino' }],
});
Dynamic path parameters

Paths with dynamic path parameters, such as /users/:id, are supported, but you need to specify the original path as a type parameter to get type validation.

import { createHttpInterceptor } from 'zimic/interceptor';

const interceptor = createHttpInterceptor<{
  '/users/:id': {
    PUT: {
      request: {
        body: {
          username: string;
        };
      };
      response: {
        204: {};
      };
    };
  };
}>({
  worker,
  baseURL: 'http://localhost:3000',
});

interceptor.get('/users/:id'); // matches any id
interceptor.get<'/users/:id'>(`/users/${1}`); // only matches id 1 (notice the original path as a type parameter)

interceptor.clear()

Clears all of the HttpRequestTracker instances created by this interceptor, including their registered responses and intercepted requests. After calling this method, the interceptor will no longer intercept any requests until new mock responses are registered.

This method is useful to reset the interceptor mocks between tests.

interceptor.clear();

HttpRequestTracker

HTTP request trackers allow declaring responses to return for matched intercepted requests. They also keep track of the intercepted requests and their responses, allowing checks about how many requests your application made and with which parameters.

When multiple trackers match the same method and path, the last created with interceptor.<method>(path) will be used.

tracker.method()

Returns the method that matches this tracker.

const tracker = interceptor.post('/users');
const method = tracker.method();
console.log(method); // 'POST'

tracker.path()

Returns the path that matches this tracker. The base URL of the interceptor is not included, but it is used when matching requests.

const tracker = interceptor.get('/users');
const path = tracker.path();
console.log(path); // '/users'

tracker.respond(declaration)

Declares a response to return for matched intercepted requests.

When the tracker matches a request, it will respond with the given declaration. The response type is statically validated against the schema of the interceptor.

Static responses
const listTracker = interceptor.get('/users').respond({
  status: 200,
  body: [{ username: 'diego-aquino' }],
});
Computed responses

A function is also supported, in case the response is dynamic:

const listTracker = interceptor.get('/users').respond((request) => {
  const username = request.searchParams.get('username');
  return {
    status: 200,
    body: [{ username }],
  };
});

The request parameter represents the intercepted request, containing useful properties such as .body, .headers, and .searchParams, which are typed based on the interceptor schema.

tracker.bypass()

Clears any declared response and intercepted requests, and makes the tracker stop matching intercepted requests. The next tracker, created before this one, that matches the same method and path will be used if present. If not, the requests to the method and path will not be intercepted.

const listTracker1 = interceptor.get('/users').respond({
  status: 200,
  body: [],
});

const listTracker2 = interceptor.get('/users').respond({
  status: 200,
  body: [{ username: 'diego-aquino' }],
});

listTracker2.bypass();

// GET requests to /users will match listTracker1 and return an empty array

tracker.requests()

Returns the intercepted requests that matched this tracker, along with the responses returned to each of them. This is useful for testing that the correct requests were made by your application.

const updateTracker = interceptor.put('/users/:id').respond((request) => {
  const newUsername = request.body.username;
  return {
    status: 200,
    body: [{ username: newUsername }],
  };
});

await fetch(`http://localhost:3000/users/${1}`, {
  method: 'PUT',
  body: JSON.stringify({ username: 'new' }),
});

const updateRequests = updateTracker.requests();
expect(updateRequests).toHaveLength(1);
expect(updateRequests[0].url).toEqual('http://localhost:3000/users/1');
expect(updateRequests[0].body).toEqual({ username: 'new' });

The return by requests are simplified objects based on the Request and Response web APIs, containing an already parsed body in .body, typed headers in .headers and typed search params in .searchParams.

If you need access to the original Request and Response objects, you can use the .raw property:

const listRequests = listTracker.requests();
console.log(listRequests[0].raw); // Request{}
console.log(listRequests[0].response.raw); // Response{}

CLI

zimic --version

Displays the current version of Zimic.

zimic --help

Displays the available commands and options.

zimic browser init <publicDirectory>

Initializes the mock service worker in a public directory.

This command is necessary when using Zimic in a browser environment. It creates a mockServiceWorker.js file in the provided public directory, which is used to intercept requests and mock responses.

If you are using Zimic mainly in tests, we recommend adding the mockServiceWorker.js to your .gitignore and adding this command to a postinstall scripts in your package.json. This ensures that the latest service worker script is being used after upgrading Zimic.


Changelog

The changelog is available on our GitHub Releases page.

Development

Interested in contributing to Zimic? View our contributing guide.