Electron Forge

A complete tool for building modern Electron applications.

Electron Forge unifies the existing (and well maintained) build tools for Electron development into a simple, easy to use package so that anyone can jump right in to Electron development.

Website | Goals | Usage | Configuration | Support | Contributing | Changelog

Getting Started

Note: Electron Forge requires Node 6 or above, plus git installed.

npm install -g electron-forge
electron-forge init my-new-app
cd my-new-app
npm start

Alternatively, if you have a more recent version of npm or yarn, you can use npx, or yarn create.

npx electron-forge init my-new-app
# or 
yarn create electron-app my-new-app
 
# then 
cd my-new-app
npm start

Project Goals

1. Starting with Electron should be as simple as a single command.
2. Developers shouldn't have to worry about babel, browserify, webpack, native module rebuilding, etc. Everything should "just work" for them out of the box.
3. Everything from creating the project to packaging the project for release should be handled by one dependency in a standard way while still offering users maximum choice and freedom.

With these goals in mind, under the hood this project uses, among others:

• electron-compile: a tool that lets you use modern and futuristic languages inside Electron without worrying about transpiling or build tooling.
• electron-rebuild: Automatically recompiles native Node.js modules against the correct Electron version.
• Electron Packager: Customizes and bundles your Electron app to get it ready for distribution.

Usage

Starting a new Project

npm install -g electron-forge
electron-forge init my-new-project

This command will generate a brand new project folder and install all your Node module dependencies, so you will be all set to go. By default we will also install the airbnb linting modules. If you want to follow the standard linting rules instead, use the --lintstyle=standard argument.

You can also start a project with your favorite framework with the --template argument. E.g. --template=react.

If you'd like to have pre-made configuration files for Travis CI and AppVeyor CI to automatically build and deploy distributables to GitHub, use the --copy-ci-files argument.

Importing an existing Project

electron-forge import existing-project-directory

Given an existing Electron project, this command will attempt to interactively navigate through the process of importing it to the Electron Forge format, so the commands listed below can be used. This includes being prompted to remove existing Electron build tools in favor of Electron Forge equivalents.

Launching your Project

electron-forge start

Any arguments after "start" will be passed through to your application when it's launched.

Packaging your Project

electron-forge package

Yes, it really is that simple. If you want to specify platform / arch, use the --platform=<platform> and --arch=<arch> arguments.

Generating a distributable for your Project

electron-forge make

This will generate platform specific distributables (installers, distribution packages, etc.) for you. By default, you can only generate distributables for your current platform. If you want to specify platform / arch, use the --platform=<platform> and --arch=<arch> arguments, but please note that some distributables are not available to be built on anything but the platform that is targeted. For example, appx (Windows Store) distributables can only be built on Windows.

Linting your Project

electron-forge lint

Publishing your Project

electron-forge publish

This will make your project and publish any generated artifacts. By default it will publish to GitHub, but you can change the publish target(s) with --target=YourTarget,YourTarget2, where the value is a comma-separated list of targets.

Config

Once you have generated a project, your package.json file will have some default forge configuration. Below is the reference structure for this config object:

{
  "make_targets": {
    "win32": ["squirrel"], // An array of win32 make targets
    "darwin": ["zip", "dmg"], // An array of darwin make targets
    "linux": ["deb", "rpm", "flatpak", "snap"] // An array of linux make targets
  },
  "electronPackagerConfig": {},
  "electronRebuildConfig": {},
  "electronWinstallerConfig": {},
  "electronInstallerDMG": {},
  "electronInstallerFlatpak": {},
  "electronInstallerDebian": {},
  "electronInstallerRedhat": {},
  "electronInstallerSnap": {}
}

Possible make targets

Configuring package

You can set electronPackagerConfig with any of the options from Electron Packager, except:

• all
• arch (use the --arch Forge command line argument instead, so it's available to all of Forge)
• asar.unpack (use asar.unpackDir instead)
• dir (use the cwd Forge command line argument instead, so it's available to all of Forge)
• electronVersion (uses the exact version specified for electron-prebuilt-compile in your devDependencies)
• out
• platform (use the --platform Forge command line argument instead, so it's available to all of Forge)
• quiet

You can set electronRebuildConfig with any of the options from Electron Rebuild, except:

• electronVersion/--version (uses the exact version specified for electron-prebuilt-compile in your devDependencies)
• arch/--arch (use the --arch Forge command line argument instead, so it's available to all of Forge)
• buildPath/--module-dir (uses your project's node_modules)

NOTE: You can also set your forge config property of your package.json to point to a JS file that exports the config object:

{
  ...
  "config": {
    "forge": "./forge.config.js"
  }
  ...
}

NOTE: If you use the JSON object then the afterCopy and afterExtract options are mapped to require calls internally, so provide a path to a file that exports your hooks and they will still run. If you use the JS file method mentioned above then you can use functions normally.

Possible publish targets

For example:

// github
{
  // Assume the GitHub repository is at https://github.com/username/repo
  "github_repository": {
    "owner": "username",
    "name": "repo"
  }
}
 
// s3
{
  "s3": {
    "accessKeyId": "<AWS_ACCESS_KEY>",
    "bucket": "my_bucket_name",
    "public": true
  }
}
 
// Electron Release Server
{
  "electronReleaseServer": {
    "baseUrl": "https://update.mysite.com",
    "username": "admin",
    "password": "no_one_will_guess_this"
  }
}
 
// Snap Store
{
  "snapStore": {
    "release": "candidate,beta"
  }
}

Custom make and publish targets

You can make your own custom targets for the make and publish targets. If you publish them as electron-forge-publisher-{name} or electron-forge-maker-{name}, they can be added to the app's devDependencies and can be specified as {name} in the make / publish targets. Publicly published third-party makers and publishers are available in the NPM registry.

The API for each target type is documented below.

API for make targets

You must export a Function that returns a Promise. Your function will be called with the following parameters.

• appDir - The directory containing the packaged application
• appName - The productName of the application
• targetArch - The target architecture of the make command
• forgeConfig - An object representing the users forgeConfig
• packageJSON - An object representing the users package.json file

Your promise must resolve with an array of the artifacts you generated.

API for publish targets

You must export a Function that returns a Promise. Your function will be called with the following keyword parameters:

• dir - The application directory
• artifacts - An array of absolute paths to artifacts to publish
• packageJSON - An object representing the user's package.json file
• forgeConfig - An object representing the user's forgeConfig
• authToken - The value of --auth-token
• tag - The value of --tag
• platform - The platform you are publishing for
• arch - The arch you are publishing for

You should use ora to indicate your publish progress.

Debugging your application on the command line

If you're using Electron 1.7 or later, you can specify the --inspect-electron flag, which will set the Electron --inspect flag with the default debugger port.

For example:

electron-forge start --inspect-electron

Debugging your application through VS Code

Debugging your Electron main process through VS Code is ridiculously easy with Forge. Simply add this as a launch config in VSCode and you're good to go.

{
  "type": "node",
  "request": "launch",
  "name": "Electron Main",
  "runtimeExecutable": "${workspaceRoot}/node_modules/.bin/electron-forge-vscode-nix",
  "windows": {
    "runtimeExecutable": "${workspaceRoot}/node_modules/.bin/electron-forge-vscode-win.cmd"
  },
  // runtimeArgs will be passed directly to your Electron application
  "runtimeArgs": [
    "foo",
    "bar"
  ],
  "cwd": "${workspaceRoot}"
}