4.1.0 • Public • Published


Node server for dynamic, fake JSON.


Dyson allows you to define JSON endpoints based on a simple path + template object:

# my-stubs/users.js
module.exports = {
  path: '/users/:userId',
  template: {
    id: params => Number(params.userId),
    name: () => faker.name.findName(),
    email: () => faker.internet.email(),
    status: (params, query) => query.status,
    lorem: true
$ dyson ./my-stubs
$ curl http://localhost:3000/users/1?status=active
  "id": 1,
  "name": "Josie Greenfelder",
  "email": "Raoul_Aufderhar@yahoo.com",
  "status": "active",
  "lorem": true

When developing client-side applications, often either static JSON files, or an actual server, backend, datastore, or API, is used. Sometimes static files are too static, and sometimes an actual server is not available, not accessible, or too tedious to set up.

This is where dyson comes in. Get a full fake server for your application up and running in minutes.

Build Status npm package dependencies npm version


  • Dynamic responses, based on
    • Request path
    • GET/POST parameters
    • Query parameters
    • Cookies
  • Dynamic HTTP status codes
  • CORS
  • Proxy (e.g. fallback to actual services)
  • Delayed responses
  • Required parameter validation
  • Includes random data generators
  • Includes dummy image generator
    • Use any external or local image service (included)
    • Supports base64 encoded image strings

Endpoint Configuration

Configure endpoints using simple objects:

module.exports = {
  path: '/user/:id',
  method: 'GET',
  template: {
    id: (params, query, body) => params.id,
    name: g.name,
    address: {
      zip: g.zipUS,
      city: g.city

The path string is the usual argument provided to Express, as in app.get(path, callback);.

The template object may contain properties of the following types:

  • A Function will be invoked with arguments (params, query, body, cookies, headers).
  • Primitives of type String, Boolean, Number, Array are returned as-is
  • An Object will be recursively iterated.
  • A Promise will be replaced with its resolved value.

Note: the template itself can also be a function returning the actual data. The template function itself is also invoked with arguments (params, query, body, cookies, headers).


The default values for the configuration objects:

module.exports = {
  cache: false,
  delay: false,
  proxy: false,
  size: () => _.random(2, 10),
  collection: false,
  callback: response.generate,
  render: response.render
  • cache: true means that multiple requests to the same path will result in the same response
  • delay: n will delay the response with n milliseconds (or between [n, m] milliseconds)
  • proxy: false means that requests to this file can be skipped and sent to the configured proxy
  • size: fn is the number of objects in the collection
  • collection: true will return a collection
  • callback: fn
    • the provided default function is doing the hard work (can be overridden)
    • used as middleware in Express
    • must set res.body and call next() to render response
  • render: fn
    • the default function to render the response (basically res.send(200, res.body);)
    • used as middleware in Express

Fake data generators

You can use anything to generate data. Here are some suggestions:

Just install the generator(s) in your project to use them in your templates:

npm install dyson-generators --save-dev


Containers can help if you need to send along some meta data, or wrap the response data in a specific way. Just use the container object, and return the data where you want it. Functions in the container object are invoked with arguments (params, query, data):

module.exports = {
  path: '/users',
  template: user.template,
  container: {
    meta: (params, query, data) => ({
      userCount: data.length
    data: {
      all: [],
      the: {
        way: {
          here: (params, query, data) => data

And an example response:

  "meta": {
    "userCount": 2
  "data": {
    "all": [],
    "the": {
      "way": {
        "here": [
            "id": 412,
            "name": "John"
            "id": 218,
            "name": "Olivia"

Combined requests

Basic support for "combined" requests is available, by means of a comma separated path fragment.

For example, a request to /user/5,13 will result in an array of the responses from /user/5 and /user/13.

The , delimiter can be configured (or disabled).

Status codes

By default, all responses are sent with a status code 200 (and the Content-Type: application/json header).

This can be overridden with your own status middleware, e.g.:

module.exports = {
  path: '/feature/:foo?',
  status: (req, res, next) => {
    if (req.params.foo === '999') {

Would result in a 404 when requesting /feature/999.


In addition to configured endpoints, dyson registers a dummy image service at /image. E.g. requesting /image/300x200 serves an image with given dimensions.

This service is a proxy to Dynamic Dummy Image Generator by Russell Heimlich.


Override the render method of the Express middleware in the endpoint definition. In the example below, depending on the existence of the callback parameter, either raw JSON response is returned or it is wrapped with the provided callback:

module.exports = {
  render: (req, res) => {
    const callback = req.query.callback;
    if (callback) {
      res.append('Content-Type', 'application/javascript');
    } else {

File Upload

Ex: return file name
formDataName = 'file'

module.exports = {
  render: (req, res) => {
    if (callback) {
      res.send({ fileName: req.files.file.name });
    } else {


If you want to run dyson over SSL you have to provide a (authority-signed or self-signed) certificate into the options.https the same way it's required for NodeJS built-in https module. Example:

const fs = require('fs');

const app = dyson.createServer({
  configDir: `${__dirname}/dummy`,
  port: 3001,
  https: {
    key: fs.readFileSync(`${__dirname}'/certs/sample.key`),
    crt: fs.readFileSync(`${__dirname}/certs/sample.crt`)

Note: if running SSL on port 443, it will require sudo privileges.


If you want dyson to support GraphQL endpoints, you can build your own logic with the render override, or use dyson-graphql. Example:

npm install dyson-graphql --save-dev
const dysonGraphQl = require('dyson-graphql');

const schema = `
  type User {
    id: Int!
    name: String!

  type Query {
    currentUser: User!

  type Mutation {
    createUser(name: String!): User!
    updateUser(id: Int!, name: String!): User!

module.exports = {
  path: '/graphql',
  method: 'POST',
  render: dysonGraphQl(schema)
    .query('currentUser', { id: 987, name: 'Jane Smart' })
    .mutation('createUser', ({ name }) => ({ id: 456, name }))
    .mutation('updateUser', ({ id, name }) => {
      if (id < 1000) {
        return { id, name };

      throw new Error("Can't update user");

Custom middleware

If you need some custom middleware before or after the endpoints are registered, dyson can be initialized programmatically. Then you can use the Express server instance (appBefore or appAfter in the example below) to install middleware before or after the dyson services are registered. An example:

const dyson = require('dyson');
const path = require('path');

const options = {
  configDir: path.join(__dirname, 'services'),
  port: 8765

const configs = dyson.getConfigurations(options);
const appBefore = dyson.createServer(options);
const appAfter = dyson.registerServices(appBefore, options, configs);

console.log(`Dyson listening at port ${options.port}`);

Dyson configuration can also be installed into any Express server:

const express = require('express');
const dyson = require('./lib/dyson');
const path = require('path');

const options = {
  configDir: path.join(__dirname, 'services')

const myApp = express();
const configs = dyson.getConfigurations(options);

dyson.registerServices(myApp, options, configs);



The recommended way to install dyson is to install it locally and put it in your package.json:

npm install dyson --save-dev

Then you can use it from scripts in package.json using e.g. npm run mocks:

  "name": "my-package",
  "version": "1.0.0",
  "scripts": {
    "mocks": "dyson mocks/"

You can also install dyson globally to start it from anywhere:

npm install -g dyson


You can put your configuration files anywhere. The HTTP method is based on:

  • The method property in the configuration itself.
  • The folder, or an ancestor folder, containing the configuration is an HTTP method. For example mocks/post/sub/endpoint.js will be an endpoint listening to POST requests.
  • Defaults to GET.
dyson [dir]

This starts the services configured in [dir] at localhost:3000.

You can also provide an alternative port number by just adding it as a second argument (e.g. dyson path/ 8181).


Project Configuration

Optionally, you can put a dyson.json file next to the configuration folders (inside [dir]). It enables to configure some behavior of dyson:

  "multiRequest": ",",
  "proxy": true,
  "proxyHost": "http://dyson.jit.su",
  "proxyPort": 8080,
  "proxyDelay": [200, 800]
  • Setting multiRequest to false disables the combined requests feature.
  • Setting bodyParserJsonLimit or bodyParserUrlencodedLimit to 1mb increases the limit to 1mb from the bodyParser's default of 100kb.
  • By default, the proxy is set to false


If you want to automatically restart dyson when you change your configuration objects, you can add nodemon as a devDependency. Say your configuration files are in the ./api folder, you can put this in your package.json:

"scripts": {
  "mocks": "dyson mocks/",
  "watch": "nodemon --watch mocks --exec dyson mocks"

Development & run tests

git clone git@github.com:webpro/dyson.git
cd dyson
npm install
npm test

Articles about dyson



Package Sidebar


npm i dyson

Weekly Downloads






Unpacked Size

27.9 kB

Total Files


Last publish


  • webpro