Skip to main content
Version: Next

Getting Started Guide

Introduction

MikroORM is a TypeScript ORM for Node.js based on Data Mapper, Unit of Work, and Identity Map patterns. In this guide, you will learn what those words mean, how to set up a simple API project, how to test it, and many more.

This Getting Started Guide was written as a step-by-step tutorial, accompanied by working StackBlitz examples and a GitHub repository with the final project. It will show you how to create a production-ready application from scratch, all the way down to a docker image you can deploy wherever you want.

The Stack

The goal of this guide is to provide show off the most important features of MikroORM as well as some of the more niche ones. It will walk you through creating a simple API for a blog, with the following technologies:

MikroORM monorepo

The ORM consists of several packages, the important ones we will be using:

  • @mikro-orm/core: the main package with the ORM code
  • @mikro-orm/cli: the CLI package, needs to be installed locally
  • @mikro-orm/sqlite: the sqlite driver package (you can use a different driver too)
  • @mikro-orm/reflection: to enable DRY entities with ts-morph reflection
  • @mikro-orm/migrations: package for managing schema migrations
  • @mikro-orm/seeder: package for seeding the database with testing data

The core and driver packages are required, the rest of this list is optional and can be a dev dependency if you wish. We will use the sqlite driver, mainly for simplicity, as it does not require any additional setup and offers a handy in-memory database, which we will use in the tests.

There are more packages, some also live outside the mikro-orm/mikro-orm monorepo, such as the @mikro-orm/nestjs or @mikro-orm/sql-highlighter - unlike the ones from the monorepo, those usually have different versioning line.

info

Full list of currently available drivers:

  • @mikro-orm/mysql
  • @mikro-orm/mariadb
  • @mikro-orm/postgresql
  • @mikro-orm/sqlite
  • @mikro-orm/better-sqlite
  • @mikro-orm/mongodb

What are we building?

We already mentioned what technologies will be used, and now more about the project. It will be a simple API for a blog, with authentication, publishing, and commenting. For that, we will use four regular entities: User, Article, Comment, and Tag. Later on, we will add one more entity, ArticleListing, a virtual entity represented by an SQL expression rather than a database table.

And the API routes description:

MethodURLDescription
POST/user/sign-upRegister new user
POST/user/sign-inLogin existing user
GET/user/profileGet your full profile info
PATCH/user/profileModify your profile
POST/articleCreate new article
GET/articleList existing articles
GET/article/:slugGet article detail
PATCH/article/:slugModify existing article
DELETE/article/:slugDelete existing article
POST/article/:slug/commentPost comment for existing article

The code will be structured into self-contained modules: user, article, and common (for shared helpers).

The app will be using Node.js 20, TypeScript 5.2, and we will build it using a modern stack with ECMAScript modules enabled.

What will we cover

Here is (an incomplete) list of features you will try going through this guide.

  • creating an app from scratch with TypeScript setup
  • folder-based discovery, ts-morph reflection, ES modules
  • request context management via middleware/fastify hook
  • entity relations, advanced entity definition (e.g. lazy scalar properties)
  • advanced type safety (e.g. OptionalProps, Reference wrapper and Loaded type)
  • events, including advanced use cases like soft-delete via onFlush event
  • basic testing via vitest
  • custom repositories
  • virtual entities
  • serialization
  • embeddables