Managing Front-end JavaScript Environment Variables

Environment variables are used in applications in order to have different behavior for different environments. For example, if an analytics tool is used in an application, it might be useful to send analytic events to different destinations depending on the environment so development and production related data doesn't get mixed together.

This article explains how the dotenv package is the basis for implementing environment variables in JavaScript and how to use environment variables with project bundlers and popular front-end frameworks. The bundlers and frameworks that will be covered in this article are the following:


dotenv

The dotenv package is by far the most popular way of setting environment variables in JavaScript. The dotenv package works by reading environment variables found in different .env files. The main .env file usually contains production variables while other .env files with different suffixes contain variables for other environments. Here are some examples:

  • .env.development (used when running an app locally or on a development server)
  • .env.staging (used when running an app on a staging server)
  • .env.test (used when running tests)

Assigning values

The way environment variables are specified in .env files as follows:

bash
1MY_VARIABLE=aCoolValue
2ANOTHER_VARIABLE=yetAnotherValue

All values assigned to environment variables are represented as strings when they are accessed in JavaScript code. That means a variable assigned as MY_VARIABLE=true will have the value of true be the string 'true' in JavaScript.

If spaces are required in the value of an environment variable, then quotes can be used around the value of the environment variable:

bash
1MY_VARIABLE="a cool value"

Reading values

This article won't discuss how to setup dotenv such that it loads the correct environment variables from the .env files since that is managed by the bundlers and frameworks that are mentioned later on in the article. However, in all the scenarios described in the article, environment variables are added to the process.env object and can, therefore, be accessed in JavaScript code using the following syntax:

bash
1process.env.MY_VARIABLE

Build time injection

Environment variables are injected into applications at build/compile time, which means that a server will need to be restarted or the project files will need to be re-built in order to see changes made to values in .env files.


Webpack

When using webpack, environment variables can be set using the DefinePlugin. DefinePlugin allows global variables to be set and made available in JavaScript code. To use DefinePlugin, add it to a webpack.config file's plugins array:

javascript
1const webpack = require("webpack");
2
3module.exports = () => ({
4 plugins: [
5 new webpack.DefinePlugin({
6 "process.env.MY_VALUE": JSON.stringify("aCoolValue")
7 })
8 ]
9});

In the above example, the environment variable can be accessed in JavaScript using process.env.MY_VALUE. It should be noted that the DefinePlugin does a search and replace of the JavaScript where it will look up any references to process.env.MY_VALUE and replace it with 'aCoolValue'. This means that MY_VALUE isn't a property on the process.env object and therefore MY_VALUE cannot be accessed by destructuring the process.env object:

javascript
1const { MY_VALUE } = process.env; // This won't work ❌

Specifying Environments

There are many ways to specify different environments with webpack. A simple way is to pass the path of the webpack.config file that corresponds with the environment you would like to build:

json
1{
2 "scripts": {
3 "build:development": "webpack --config webpack.config.development.js",
4 "build:staging": "webpack --config webpack.config.staging.js",
5 "build:production": "webpack --config webpack.config.production.js"
6 }
7}

Resources

Webpack DefinePlugin documentation

Using dotenv in combination with DefinePlugin


Parcel

Parcel uses dotenv to manage environment variables. All that is required is to create .env files for different environments and then those variables will be made accessible in JavaScript code using the process.env.VARIABLE_NAME syntax.

Specifying Environments

When running the normal parcel serve command (e.g. parcel index.html), the environment (specified by the NODE_ENV environment variable) will be development by default. To specify an environment other than development while using the parcel serve command, you can specify the environment when running the parcel serve command:

bash
1NODE_ENV=staging parcel index.html

When running the parcel build command (e.g. parcel build index.html), the environment will be production by default. Again, you can specify a specific environment when running the build command:

bash
1NODE_ENV=staging parcel build index.html

Order of priority/inheritance

Environment variables can be taken from multiple .env files. For example, all environments will inherit environment variables from the main .env file. If in a development environment, you can inherit values from 4 different files: .env.development.local, .env.development, .env.local, and .env. If the same variables are found in multiple .env files, then the value for the environment variable will be taken with the .env file with the highest priority. Here is the order of priority, from highest to lowest:

  • .env.${NODE_ENV}.local
  • .env.${NODE_ENV}
  • .env.local
  • .env

Note: The test environment ignores all local .env files. This is done because the tests should produce the same result for everyone.

Source control

You should commit all your .env files to source control with the exception of .env*.local files. The local .env files should only be used to tweak your app's configurations when running the app locally.

Resources

Parcel environment variables

Parcel dotenv configuration (repo code)


Create React App

Create React App uses both dotenv and webpack's DefinePlugin to manage environment variables. Just like with parcel, environment variables need to be defined in .env files at the same level as the project's package.json file.

Naming requirements

Environment variables need to be prefixed by REACT_APP_ in order for the variables to be accessible on the process.env object in JavaScript.

bash
1REACT_APP_MY_VARIABLE=testing123

Specifying environments

Create React App does not allow to change the value of the NODE_ENV environment variable. The npm start command will set the NODE_ENV to development, the npm test command will set the NODE_ENV to test, and the npm run build command sets the NODE_ENV to production.

Given that the NODE_ENV is set for you and that the value for NODE_ENV is used to reconcile the correct .env file, the following .env files can be used:

  • .env
  • .env.local (loaded for all environments except test)
  • .env.development, .env.test, .env.production
  • .env.development.local, .env.test.local, .env.production.local

Order of priority/inheritance

The order of priority and inheritance is exactly the same as that described for the parcel bundler:

  • .env.${NODE_ENV}.local
  • .env.${NODE_ENV}
  • .env.local
  • .env

Additional environments

To use environment variables for environments other than development, test, and production, you can create additional .env files and load the correct .env file using env-cmd.

To take a staging environment as an example:

  • Create a .env.staging file and add environment variables to the file
  • Add env-cmd as a project dependency (npm install env-cmd --save)
  • Create script commands for the staging environment
json
1{
2 "scripts": {
3 "start:staging": "env-cmd .env.staging npm start",
4 "build:staging": "env-cmd .env.staging npm run build"
5 }
6}
  • Run the start:staging or build:staging command to start a local staging environment or to build the staging environment bundle

It's important to note that the NODE_ENV will still be set to development when running the npm start command and the NODE_ENV will be set to production when running the npm run build command, so environment variables can still be loaded from either .env.development or .env.production depending on the command used.

For example, running the start:staging command from above would load environment variables from the following files (in order of priority):

  • .env.staging
  • .env.development.local
  • .env.development
  • .env.local
  • .env

Source control

Just like with Parcel, you should commit all your .env files to source control with the exception of .env*.local files. The local .env files should only be used to tweak your app's configurations when running the app locally.

Resources

CRA environment variable documentation

CRA dotenv configuration (repo code)

CRA webpack configuration (repo code)


Gatsby

Gatsby handles environment variables in a unique way compared to Create React App and Parcel. Gatsby uses both dotenv and webpack's DefinePlugin to manage environment variables.

By default, Gatsby supports a development and a production environment, so environment variables should be added to either a .env.development or a .env.production file. Gatsby does not support .env files that do not contain a suffix.

Naming requirements

Any environment variables that are used in the front-end code should be prefixed with GATSBY_ in order for the environment variables to work properly in a deployed environment.

If an environment variable is not used in the front-end code but rather used at some point in the Gatsby build process, it does not need to be prefixed by GATSBY_ when the environment variable is defined in .env files.

See the "Source control" section below for some exceptions to these naming requirements.

Specifying environments

When running the gatsby develop command, that will set the NODE_ENV to development and environment variables will be taken from .env.development. When building the app using gatsby build, the NODE_ENV will be set to production and environment variables will be taken from .env.production.

If you'd like to have more than a development and a production environment, you can do so by specifying a value for GATSBY_ACTIVE_ENV when running one of the gatsby commands. For example, if you were to have a .env.staging file, you would be able to access the variables from that file by running one of the following commands:

bash
1GATSBY_ACTIVE_ENV=staging gatsby develop
2GATSBY_ACTIVE_ENV=staging gatsby build

Note: if using GATSBY_ACTIVE_ENV to select a different .env file, then the environment variables from .env.development and .env.production will be ignored.

Source control

Gatsby recommends that you don't commit your .env files to source control, but rather add environment variables to your app's hosting provider's deployment tool. This is because sometimes API keys are required to perform certain operations during the Gatsby build step (which is done in a backend-end environment). Because you don't want these private API keys exposed to the public through source control, these API keys should be stored in .env files that are not committed to source control.

If you do decide to commit your .env files to source control and make those available on your hosted environment, you do not need to prefix any of your environment variables by GATSBY_, even if those environment variables are used on the front-end.

Resources

Gatsby environment variable documentation

Gatsby dotenv config (repo code)

Gatsby webpack config (repo code)


A note on environment variables and source control

Almost all the articles I've read on environment variables say to never commit your .env files to source control, so I've been terrified to do so for so long. However, this is bad advice. If all your environment variables will be embedded in your front-end JavaScript code, then there is no harm in committing your .env files to source control. Any environment variable used in your front-end code will be viewable in the browser if a user decides to inspect the browser's source code.

In fact, I would recommend committing these files in source control as this will make the installation process for a project much simpler for all developers. It's even OK to commit API keys saved in your environment variables as long as you're sure those API keys are being used in your front-end code and that is how they are intended to be used. For example, if you'd like to embed Google Maps on a web page, it's required to use a Google Maps API key to do that. That API key has no choice but to be exposed to the public since it will be part of an iframe's src value.

You might think that having an API key exposed to the public is bad since anyone will be able to use that API key and then potentially max our your API call limits. However, API keys that need to be exposed to the front-end (like the Google Maps API key mentioned above), have ways to restrict the usage of the API keys by restricting usage to certain IP addresses, referrer URLs, and mobile apps. You can read Google's recommendation for API Key Best Practices for more details.

With all that being said, if you are using API keys for back-end code (like what could be the case when building a Gatsby site), then the API keys should be kept in .env files and those files should not be committed to source control.


Hopefully, this article gives you a better idea of when, why, and how to use environment variables in your front-end JavaScript applications. Let me know if you have other useful tips related to environment variables.

Interested in learning how environment variables should be configured on the back-end with Node.js? Read this excellent article by John Papa.



Robert Cooper's big ol' head

Hey, I'm Robert Cooper and I write articles related to web development. If you find these articles interesting, follow me on Twitter to get more bite-sized content related to web development.