Skip to content
GraphQL Joker is the ultimate GraphQL scaffolding tool.
JavaScript TypeScript
Branch: master
Clone or download

GraphQL Joker

NPM version Build Status Test Coverage Dependency Status License PR Welcome

GraphQL Joker is the ultimate GraphQL scaffolding tool.

It automates coding process to save your precious time, enhance your work and life experience. In other words, GraphQL Joker write code for you with commands you specified.

With GraphQL Joker, you can create a full-fledged backend server with your complex app logic and running API in less than 3 minutes.



When working on GraphQL projects, we need to define database schema and GraphQL twice. We need to create resolvers for the standardized API. A lot of copying and pasting are going on. It's not elegant to copy code around and find-replace all occurrences. It's also error prone. And sometimes causing unnoticeable errors which wastes time.

Wouldn't be nice if we could have a tool just like Ruby on Rails' scaffold tool to generate code for us?

Design Concept

GraphQL Joker is designed to provide features at least Ruby on Rails scaffold tool has. Aka, generate boilerplate business logic code as much as possible for you.

However, unlike Ruby on Rails, GraphQL Joker is not a full-fledged framework and will never provide a framework for you. It focus on business logic generation. Although GraphQL Joker also has project generation feature, it's trying to hook up the industry standard and battle-tested libraries and components together for you. And it's configurable. To split features into small core chunks and make them combinable and adaptable is a good practice and especially popular in node.js ecosystem, GraphQL Joker embraces this practice. That's what makes GraphQL Joker outstanding and what makes GraphQL Joker really a flexible and configurable scaffolding tool.


GraphQL Joker is a general command line tool, thus you should install it globally.

npm install -g graphql-joker

Create an GraphQL Project

To create an GraphQL project, use joker app command.

joker app my-new-app

This will generate your app in 'my-new-app' folder. If you don't specify app name, the app will be created at your current working directory.


  • --port On which port this app is listening on.
  • --git-init Automatically run 'git init' after project generated.
  • --skip-install Do not install dependencies.
  • --eslint-config Changing the default eslint config.
  • --main Changing the entry filename.

To change default eslint config being used:

joker app my-new-app --eslint-config=your-config

To automatically run git init:

joker app my-new-app --git-init

Generate Resources

API resource generation is the core feature of GraphQL Joker. It's syntax is rather simple and extensible. It follows this basic style:

joker resource ModelName[/optionalPluralVariableName] \
primitiveField[[:Type[typeModifiers]]:defaultValue]... \

This arguments specification is obscure to see. Let's see some examples.

Let's say you have a model named user, and user has a name, an age and also a list of posts. And you have a model named post, it has title, content and author. Just type like this:

joker resource User name:String age:Int posts:[Post]:author
joker resource Post title:String content:String author:User

Here we specified our first model 'User', with following fields:

  • name which is a String
  • age which is an Int
  • posts which is a list of Posts through foreign key named author

We defined our second model named 'Post', with following fields:

  • title which is a String
  • content which is also a String
  • author which references to User

This creates six files in total, three for User and three for Post. The three files are mongoose model, GraphQL schema and GraphQL resolver.

The autogenerated models/User.js looks like this:

const mongoose = require('mongoose');
const { Schema } = mongoose;

const userSchema = new Schema({
  name: String,
  age: Number
}, {
  timestamps: true,
  collection: 'users'

module.exports = mongoose.model('User', userSchema);

The autogenerated schemas/User.gql looks like this:

type User {
  _id: ID!
  name: String
  age: Int
  posts: [Post]
  createdAt: Date
  updatedAt: Date

input UserInput {
  name: String
  age: Int

type Query {
  user(_id: ID!): User
  users: [User]

type Mutation {
  createUser(input: UserInput): User
  updateUser(_id: ID!, input: UserInput): User
  deleteUser(_id: ID!): User

The autogenerated resolvers/User.js looks like this:

module.exports = {
  User: {
    async posts(root, _, { Post }) {
      return await Post.find({ author: root._id });
  Query: {
    async user(root, { _id }, { User }) {
      return await User.findById(_id);
    async users(root, { _ }, { User }) {
      return await User.find();
  Mutation: {
    async createUser(root, { input }, { User }) {
      return await User.create(input);
    async updateUser(root, { _id, input }, { User }) {
      return await (await User.findById(_id)).set(input).save();
    async deleteUser(root, { _id }, { User }) {
      return await (await User.findById(_id)).remove();

Besides your schema definition, 5 API are created for you. Those are:

  • users query all users
  • user query a user by id
  • createUser create a new user
  • updateUser modify an existing user
  • deleteUser delete an existing user

Now you can CRUD your resources through API.

Primitive Types

GraphQL Joker supports a wide range of primitive types:

  • String string type
  • Int integer type
  • Float float type
  • Boolean bool type
  • Date date type
  • Enum enum type, the type specifier has a different syntax
  • File upload typem the type specifier has a different syntax
  • Mixed mixed type includes string, int, float, boolean, date, array and objects

When you are defining a field with type mentioned above, GraphQL Joker will treat them as primitive types. When you refer to a type that is not included in the list, GraphQL Joker will treat it as a referecing to another model.

joker resource User disabled:Boolean name:String description:Mixed spouse:User

In the above example, obviously disabled, name and description are primitive types. spouse is a reference type which references to User.

Array Type

Surround a type with a pair of [], you get an array of that type, for example:

joker resource User spouse:User friends:[User] favoriteSayings:[String]

The field friends is an array of Users. And the field favoriteSayings is an array of Strings.

Reference Types

There are several ways to implement your own reference types.


The simplest case is one-to-one relation ship.

joker resource User address:Address
joker resource Address user:User:address

In this case, we save the reference into user model, and on address model, we use the foreign key on user model to fetch the user value.


We have two ways to implement this relationship.

joker resource User posts:[Post]:owner
joker resource Post user:User:owner

This is the most common case. We save the reference on the 'many' side, and fetch on the 'many' side model.

joker resource User posts:[Post]
joker resource Post user:User:[posts]

In this case, we are saving the references on the 'one' side, and on 'many' side, we use a pair of [] to indicate it's an array. Be careful of performance when you are doing this way.


In simple cases, we can just do like this.

joker resource User courses:[Course]
joker resource Course users:[User]:[courses]

If there are tons of records, then you may want to use association table.

joker resource Favorite user:User course:Course
joker resource User courses:[Course]:Favorite
joker resource Course users:[User]:Favorite

In this case, we specified a relationship that is have many ... through ...

Upload Type

To create an uploading field, use ...Uploader as type name. See the following example:

joker resource User avatar:AvatarUploader

To create an uploader, see Generate Uploader

Type Modifiers

In the real world practices, fields should be validated. For example, You may want a user's email to match designated format and to be required and unique. You can specify type modifiers.

joker resource User 'email:String/.*@.*\..*/!$'

In the above example, /.*@.*\..*/ means that this field matches this regexp, ! means required, and $ means unique.

Existing type modifiers includes:

  • ! required
  • ^ index
  • $ unique
  • /regexp/ string only, matches the regexp or not
  • <=n max for number types, maxlength for string type
  • >=n min for number types, minlength for string type

Default Values

You can specify default value to a primitive field with the following syntax.

joker resource Post 'title:String!:Untitled' 'lastUpdate:Date!:``'

Here, title's default value is 'Untitled', and lastUpdate's default value is It's a calculated default value, so surround with a pair of back ticks.

Nested Structure

To create nested structure, use the following syntax:

joker resource User posts:[{ title:String content:String comments:[{ \
commenter:User content:String }] }] email:String password:String settings:{ \
sms:Boolean email:Boolean pushNotification:Boolean }

Specify type as { or [{, you are going into a nested context. All field defined after this goes into the nested structure. Use plain } and }] tokens to jump out the nesting context.


To create enum fields, use enum syntax like this:

joker resource User 'gender:Enum(male,female)!'

Reusable Nestables

GraphQL Joker supports reusable nestables and referencing them.

joker nestable Address line1:String line2:String country:String region:String
joker resource User address:addressSchema name:String

Specify the lowercase nestable name append by 'Schema', joker will treat the type as a subschema reference.

Destroy Resources

If you mistakenly generated something or you spell something wrongly, use the 'destroy' command to delete the autogenerated files. Just append destroy with the original command, it automatically destroys the generated content.

joker destroy resource User name:String

Generate Uploader

To generate an uploader, use joker uploader command.

joker uploader FileUploader extends AliOSSUploader bucket=your-bucket-name region=your-region

This generates an base file uploader for you.

Integrate with Existing Project

GraphQL Joker is designed to be a generic tool. It does not require a project to be a GraphQL Joker project.

Customize GraphQL Joker Behavior

Create a file called .jokerrc.json in project's root directory. And filling it like this:

  "schemaDir": "graphql",
  "resolverDir": "graphql",
  "test": false

GraphQL Joker will generate schema files and resolver files into graphql directory, and will not generate unit tests.

Issues and Helps

GraphQL Joker is not mature yet. If you find anything uncomfortable or confuses you. Any discuss, issue and pull request are welcome.


GraphQL Joker is an ambitious project and it still has a long way to go.

  • Version 0.8
    • Basic sequelize support
  • Version 0.9
    • CLI user experience
    • use eslint to transform user generated code if available
    • configurability
    • dependencies reliability
  • Version 0.10
    • query filter, sorting and pagination feature


The GNU General Public License v3.0.

You can’t perform that action at this time.