package.json
Select CLI Version:
See Details
Table of contents
Description
This document is all you need to know about what's required in your package.json file. It must be actual JSON, not just a JavaScript object literal.
A lot of the behavior described in this document is affected by the config settings described inconfig
.
name
If you plan to publish your package, themost important things in your package.json are the name and version fields as they will be required. The name and version together form an identifier that is assumed to be completely unique. If you don't plan to publish your package, the name and version fields are optional. The name field contains your package name.
Some rules:
- The name must be less than or equal to 214 characters. This includes the scope for scoped packages.
- The names of scoped packages can begin with a dot or an underscore. This is not permitted without a scope.
- New packages must not have uppercase letters in the name.
- The name ends up being part of a URL, an argument on the command line, and a folder name. Therefore, the name can't contain any non-URL-safe characters.
Some tips:
- Don't use the same name as a core Node module.
- Don't put "js" or "node" in the name. It's assumed that it's js, since you're writing a package.json file, and you can specify the engine using the "engines" field. (See below.)
- The name will probably be passed as an argument to require(), so it should be something short, but also reasonably descriptive.
- You may want to check the npm registry to see if there's something by that name already, before you get too attached to it.https://www.npmjs.com/
A name can be optionally prefixed by a scope, e.g.@myorg/mypackage
. Seescope
for more detail.
version
Changes to the package should come along with changes to the version. You can show developers how much they need to adjust on a new update by usingsemantic versioning
Version must be parseable bynode-semver, which is bundled with npm as a dependency. (npm install semver
to use it yourself.)
description
Put a description in it. It's a string. This helps people discover your package, as it's listed innpm search
.
keywords
Put keywords in it. It's an array of strings. This helps people discover your package as it's listed innpm search
.
homepage
The url to the project homepage.
Example:
"homepage":"https://github.com/owner/project#readme"
bugs
The url to your project's issue tracker and / or the email address to which issues should be reported. These are helpful for people who encounter issues with your package.
It should look like this:
{"url":"https://github.com/owner/project/issues","email":"project@hostname.com"}
You can specify either one or both values. If you want to provide only a url, you can specify the value for "bugs" as a simple string instead of an object.
If a url is provided, it will be used by thenpm bugs
command.
license
You should specify a license for your package so that people know how they are permitted to use it, and any restrictions you're placing on it.
If you're using a common license such as BSD-2-Clause or MIT, add a current SPDX license identifier for the license you're using, like this:
{"license":"BSD-3-Clause"}
You can checkthe full list of SPDX license IDs. Ideally you should pick one that isOSI approved.
If your package is licensed under multiple common licenses, use anSPDX license expression syntax version 2.0 string, like this:
{"license":"(ISC OR GPL-3.0)"}
If you are using a license that hasn't been assigned an SPDX identifier, or if you are using a custom license, use a string value like this one:
{"license":"SEE LICENSE IN <filename>"}
Then include a file named<filename>
at the top level of the package.
Some old packages used license objects or a "licenses" property containing an array of license objects:
// Not valid metadata{"license":{"type":"ISC","url":"https://opensource.org/licenses/ISC"}}// Not valid metadata{"licenses":[{"type":"MIT","url":"https://www.opensource.org/licenses/mit-license.php"},{"type":"Apache-2.0","url":"https://opensource.org/licenses/apache2.0.php"}]}
Those styles are now deprecated. Instead, use SPDX expressions, like this:
{"license":"ISC"}
{"license":"(MIT OR Apache-2.0)"}
Finally, if you do not wish to grant others the right to use a private or unpublished package under any terms:
{"license":"UNLICENSED"}
Consider also setting"private": true
to prevent accidental publication.
people fields: author, contributors
The "author" is one person. "contributors" is an array of people. A "person" is an object with a "name" field and optionally "url" and "email", like this:
{"name":"Barney Rubble","email":"b@rubble.com","url":"http://barnyrubble.tumblr.com/"}
Or you can shorten that all into a single string, and npm will parse it for you:
{"author":"Barney Rubble <b@rubble.com> (http://barnyrubble.tumblr.com/)"}
Both email and url are optional either way.
npm also sets a top-level "maintainers" field with your npm user info.
funding
You can specify an object containing a URL that provides up-to-date information about ways to help fund development of your package, or a string URL, or an array of these:
{"funding":{"type":"individual","url":"http://example.com/donate"},"funding":{"type":"patreon","url":"https://www.patreon.com/my-account"},"funding":"http://example.com/donate","funding":[{"type":"individual","url":"http://example.com/donate"},"http://example.com/donateAlso",{"type":"patreon","url":"https://www.patreon.com/my-account"}]}
Users can use thenpm fund
subcommand to list thefunding
URLs of all dependencies of their project, direct and indirect. A shortcut to visit each funding url is also available when providing the project name such as:npm fund <projectname>
(when there are multiple URLs, the first one will be visited)
files
The optionalfiles
field is an array of file patterns that describes the entries to be included when your package is installed as a dependency. File patterns follow a similar syntax to.gitignore
, but reversed: including a file, directory, or glob pattern (*
,**/*
, and such) will make it so that file is included in the tarball when it's packed. Omitting the field will make it default to["*"]
, which means it will include all files.
Some special files and directories are also included or excluded regardless of whether they exist in thefiles
array (see below).
You can also provide a.npmignore
file in the root of your package or in subdirectories, which will keep files from being included. At the root of your package it will not override the "files" field, but in subdirectories it will. The.npmignore
file works just like a.gitignore
. If there is a.gitignore
file, and.npmignore
is missing,.gitignore
's contents will be used instead.
Files included with the "package.json#files" fieldcannot be excluded through.npmignore
or.gitignore
.
Certain files are always included, regardless of settings:
package.json
README
LICENSE
/LICENCE
- The file in the "main" field
README
&LICENSE
can have any case and extension.
Conversely, some files are always ignored:
.git
CVS
.svn
.hg
.lock-wscript
.wafpickle-N
.*.swp
.DS_Store
._*
npm-debug.log
.npmrc
node_modules
config.gypi
*.orig
package-lock.json
(usenpm-shrinkwrap.json
if you wish it to be published)
main
The main field is a module ID that is the primary entry point to your program. That is, if your package is namedfoo
, and a user installs it, and then doesrequire("foo")
, then your main module's exports object will be returned.
This should be a module relative to the root of your package folder.
For most modules, it makes the most sense to have a main script and often not much else.
Ifmain
is not set it defaults toindex.js
in the package's root folder.
browser
If your module is meant to be used client-side the browser field should be used instead of the main field. This is helpful to hint users that it might rely on primitives that aren't available in Node.js modules. (e.g.window
)
bin
A lot of packages have one or more executable files that they'd like to install into the PATH. npm makes this pretty easy (in fact, it uses this feature to install the "npm" executable.)
To use this, supply abin
field in your package.json which is a map of command name to local file name. When this package is installed globally, that file will be linked where global bins go so it is available to run by name. When this package is installed as a dependency in another package, the file will be linked where it will be available to that package either directly bynpm exec
or by name in other scripts when invoking them vianpm run-script
.
For example, myapp could have this:
{"bin":{"myapp":"./cli.js"}}
So, when you install myapp, it'll create a symlink from thecli.js
script to/usr/local/bin/myapp
.
If you have a single executable, and its name should be the name of the package, then you can just supply it as a string. For example:
{"name":"my-program","version":"1.2.5","bin":"./path/to/program"}
would be the same as this:
{"name":"my-program","version":"1.2.5","bin":{"my-program":"./path/to/program"}}
Please make sure that your file(s) referenced inbin
starts with#!/usr/bin/env node
, otherwise the scripts are started without the node executable!
Note that you can also set the executable files usingdirectories.bin.
Seefolders for more info on executables.
man
Specify either a single file or an array of filenames to put in place for theman
program to find.
If only a single file is provided, then it's installed such that it is the result fromman <pkgname>
, regardless of its actual filename. For example:
{"name":"foo","version":"1.2.3","description":"A packaged foo fooer for fooing foos","main":"foo.js","man":"./man/doc.1"}
would link the./man/doc.1
file in such that it is the target forman foo
If the filename doesn't start with the package name, then it's prefixed. So, this:
{"name":"foo","version":"1.2.3","description":"A packaged foo fooer for fooing foos","main":"foo.js","man":["./man/foo.1","./man/bar.1"]}
will create files to doman foo
andman foo-bar
.
Man files must end with a number, and optionally a.gz
suffix if they are compressed. The number dictates which man section the file is installed into.
{"name":"foo","version":"1.2.3","description":"A packaged foo fooer for fooing foos","main":"foo.js","man":["./man/foo.1","./man/foo.2"]}
will create entries forman foo
andman 2 foo
directories
The CommonJSPackages spec details a few ways that you can indicate the structure of your package using adirectories
object. If you look atnpm's package.json, you'll see that it has directories for doc, lib, and man.
In the future, this information may be used in other creative ways.
directories.bin
If you specify abin
directory indirectories.bin
, all the files in that folder will be added.
Because of the way thebin
directive works, specifying both abin
path and settingdirectories.bin
is an error. If you want to specify individual files, usebin
, and for all the files in an existingbin
directory, usedirectories.bin
.
directories.man
A folder that is full of man pages. Sugar to generate a "man" array by walking the folder.
repository
Specify the place where your code lives. This is helpful for people who want to contribute. If the git repo is on GitHub, then thenpm docs
command will be able to find you.
Do it like this:
{"repository":{"type":"git","url":"https://github.com/npm/cli.git"}}
The URL should be a publicly available (perhaps read-only) url that can be handed directly to a VCS program without any modification. It should not be a url to an html project page that you put in your browser. It's for computers.
For GitHub, GitHub gist, Bitbucket, or GitLab repositories you can use the same shortcut syntax you use fornpm install
:
{"repository":"npm/npm","repository":"github:user/repo","repository":"gist:11081aaa281","repository":"bitbucket:user/repo","repository":"gitlab:user/repo"}
If thepackage.json
for your package is not in the root directory (for example if it is part of a monorepo), you can specify the directory in which it lives:
{"repository":{"type":"git","url":"https://github.com/facebook/react.git","directory":"packages/react-dom"}}
scripts
The "scripts" property is a dictionary containing script commands that are run at various times in the lifecycle of your package. The key is the lifecycle event, and the value is the command to run at that point.
Seescripts
to find out more about writing package scripts.
config
A "config" object can be used to set configuration parameters used in package scripts that persist across upgrades. For instance, if a package had the following:
{"name":"foo","config":{"port":"8080"}}
It could also have a "start" command that referenced thenpm_package_config_port
environment variable.
dependencies
Dependencies are specified in a simple object that maps a package name to a version range. The version range is a string which has one or more space-separated descriptors. Dependencies can also be identified with a tarball or git URL.
Please do not put test harnesses or transpilers or other "development" time tools in yourdependencies
object. SeedevDependencies
, below.
Seesemver for more details about specifying version ranges.
version
Must matchversion
exactly>version
Must be greater thanversion
>=version
etc<version
<=version
~version
"Approximately equivalent to version" Seesemver^version
"Compatible with version" Seesemver1.2.x
1.2.0, 1.2.1, etc., but not 1.3.0http://...
See 'URLs as Dependencies' below*
Matches any version""
(just an empty string) Same as*
version1 - version2
Same as>=version1 <=version2
.range1 || range2
Passes if either range1 or range2 are satisfied.git...
See 'Git URLs as Dependencies' belowuser/repo
See 'GitHub URLs' belowtag
A specific version tagged and published astag
Seenpm dist-tag
path/path/path
SeeLocal Paths below
For example, these are all valid:
{"dependencies":{"foo":"1.0.0 - 2.9999.9999","bar":">=1.0.2 <2.1.2","baz":">1.0.2 <=2.3.4","boo":"2.0.1","qux":"<1.0.0 || >=2.3.1 <2.4.5 || >=2.5.2 <3.0.0","asd":"http://asdf.com/asdf.tar.gz","til":"~1.2","elf":"~1.2.3","two":"2.x","thr":"3.3.x","lat":"latest","dyl":"file:../dyl"}}
URLs as Dependencies
You may specify a tarball URL in place of a version range.
This tarball will be downloaded and installed locally to your package at install time.
Git URLs as Dependencies
Git urls are of the form:
<protocol>://[<user>[:<password>]@]<hostname>[:<port>][:][/]<path>[#<commit-ish> | #semver:<semver>]
<protocol>
is one ofgit
,git+ssh
,git+http
,git+https
, orgit+file
.
If#<commit-ish>
is provided, it will be used to clone exactly that commit. If the commit-ish has the format#semver:<semver>
,<semver>
can be any valid semver range or exact version, and npm will look for any tags or refs matching that range in the remote repository, much as it would for a registry dependency. If neither#<commit-ish>
or#semver:<semver>
is specified, then the default branch is used.
Examples:
git+ssh://git@github.com:npm/cli.git#v1.0.27git+ssh://git@github.com:npm/cli#semver:^5.0git+https://isaacs@github.com/npm/cli.gitgit://github.com/npm/cli.git#v1.0.27
When installing from agit
repository, the presence of certain fields in thepackage.json
will cause npm to believe it needs to perform a build. To do so your repository will be cloned into a temporary directory, all of its deps installed, relevant scripts run, and the resulting directory packed and installed.
This flow will occur if your git dependency usesworkspaces
, or if any of the following scripts are present:
build
prepare
prepack
preinstall
install
postinstall
If your git repository includes pre-built artifacts, you will likely want to make sure that none of the above scripts are defined, or your dependency will be rebuilt for every installation.
GitHub URLs
As of version 1.1.65, you can refer to GitHub urls as just "foo": "user/foo-project". Just as with git URLs, acommit-ish
suffix can be included. For example:
{"name":"foo","version":"0.0.0","dependencies":{"express":"expressjs/express","mocha":"mochajs/mocha#4727d357ea","module":"user/repo#feature\/branch"}}
Local Paths
As of version 2.0.0 you can provide a path to a local directory that contains a package. Local paths can be saved usingnpm install -S
ornpm install --save
, using any of these forms:
../foo/bar~/foo/bar./foo/bar/foo/bar
in which case they will be normalized to a relative path and added to yourpackage.json
. For example:
{"name":"baz","dependencies":{"bar":"file:../foo/bar"}}
This feature is helpful for local offline development and creating tests that require npm installing where you don't want to hit an external server, but should not be used when publishing packages to the public registry.
note: Packages linked by local path will not have their own dependencies installed whennpm install
is ran in this case. You must runnpm install
from inside the local path itself.
devDependencies
If someone is planning on downloading and using your module in their program, then they probably don't want or need to download and build the external test or documentation framework that you use.
In this case, it's best to map these additional items in adevDependencies
object.
These things will be installed when doingnpm link
ornpm install
from the root of a package, and can be managed like any other npm configuration param. Seeconfig
for more on the topic.
For build steps that are not platform-specific, such as compiling CoffeeScript or other languages to JavaScript, use theprepare
script to do this, and make the required package a devDependency.
For example:
{"name":"ethopia-waza","description":"a delightfully fruity coffee varietal","version":"1.2.3","devDependencies":{"coffee-script":"~1.6.3"},"scripts":{"prepare":"coffee -o lib/ -c src/waza.coffee"},"main":"lib/waza.js"}
Theprepare
script will be run before publishing, so that users can consume the functionality without requiring them to compile it themselves. In dev mode (ie, locally runningnpm install
), it'll run this script as well, so that you can test it easily.
peerDependencies
In some cases, you want to express the compatibility of your package with a host tool or library, while not necessarily doing arequire
of this host. This is usually referred to as aplugin. Notably, your module may be exposing a specific interface, expected and specified by the host documentation.
For example:
{"name":"tea-latte","version":"1.3.5","peerDependencies":{"tea":"2.x"}}
This ensures your packagetea-latte
can be installedalong with the second major version of the host packagetea
only.npm install tea-latte
could possibly yield the following dependency graph:
├── tea-latte@1.3.5└── tea@2.2.0
In npm versions 3 through 6,peerDependencies
were not automatically installed, and would raise a warning if an invalid version of the peer dependency was found in the tree. As of npm v7, peerDependenciesare installed by default.
Trying to install another plugin with a conflicting requirement may cause an error if the tree cannot be resolved correctly. For this reason, make sure your plugin requirement is as broad as possible, and not to lock it down to specific patch versions.
Assuming the host complies withsemver, only changes in the host package's major version will break your plugin. Thus, if you've worked with every 1.x version of the host package, use"^1.0"
or"1.x"
to express this. If you depend on features introduced in 1.5.2, use"^1.5.2"
.
peerDependenciesMeta
When a user installs your package, npm will emit warnings if packages specified inpeerDependencies
are not already installed. ThepeerDependenciesMeta
field serves to provide npm more information on how your peer dependencies are to be used. Specifically, it allows peer dependencies to be marked as optional.
For example:
{"name":"tea-latte","version":"1.3.5","peerDependencies":{"tea":"2.x","soy-milk":"1.2"},"peerDependenciesMeta":{"soy-milk":{"optional":true}}}
Marking a peer dependency as optional ensures npm will not emit a warning if thesoy-milk
package is not installed on the host. This allows you to integrate and interact with a variety of host packages without requiring all of them to be installed.
bundleDependencies
This defines an array of package names that will be bundled when publishing the package.
In cases where you need to preserve npm packages locally or have them available through a single file download, you can bundle the packages in a tarball file by specifying the package names in thebundleDependencies
array and executingnpm pack
.
For example:
If we define a package.json like this:
{"name":"awesome-web-framework","version":"1.0.0","bundleDependencies":["renderized","super-streams"]}
we can obtainawesome-web-framework-1.0.0.tgz
file by runningnpm pack
. This file contains the dependenciesrenderized
andsuper-streams
which can be installed in a new project by executingnpm install awesome-web-framework-1.0.0.tgz
. Note that the package names do not include any versions, as that information is specified independencies
.
If this is spelled"bundledDependencies"
, then that is also honored.
Alternatively,"bundleDependencies"
can be defined as a boolean value. A value oftrue
will bundle all dependencies, a value offalse
will bundle none.
optionalDependencies
If a dependency can be used, but you would like npm to proceed if it cannot be found or fails to install, then you may put it in theoptionalDependencies
object. This is a map of package name to version or url, just like thedependencies
object. The difference is that build failures do not cause installation to fail. Runningnpm install --omit=optional
will prevent these dependencies from being installed.
It is still your program's responsibility to handle the lack of the dependency. For example, something like this:
try{var foo=require("foo");var fooVersion=require("foo/package.json").version;}catch(er){foo=null;}if(notGoodFooVersion(fooVersion)){foo=null;}// .. then later in your program ..if(foo){foo.doFooThings();}
Entries inoptionalDependencies
will override entries of the same name independencies
, so it's usually best to only put in one place.
overrides
If you need to make specific changes to dependencies of your dependencies, for example replacing the version of a dependency with a known security issue, replacing an existing dependency with a fork, or making sure that the same version of a package is used everywhere, then you may add an override.
Overrides provide a way to replace a package in your dependency tree with another version, or another package entirely. These changes can be scoped as specific or as vague as desired.
To make sure the packagefoo
is always installed as version1.0.0
no matter what version your dependencies rely on:
{"overrides":{"foo":"1.0.0"}}
The above is a short hand notation, the full object form can be used to allow overriding a package itself as well as a child of the package. This will causefoo
to always be1.0.0
while also makingbar
at any depth beyondfoo
also1.0.0
:
{"overrides":{"foo":{".":"1.0.0","bar":"1.0.0"}}}
To only overridefoo
to be1.0.0
when it's a child (or grandchild, or great grandchild, etc) of the packagebar
:
{"overrides":{"bar":{"foo":"1.0.0"}}}
Keys can be nested to any arbitrary length. To overridefoo
only when it's a child ofbar
and only whenbar
is a child ofbaz
:
{"overrides":{"baz":{"bar":{"foo":"1.0.0"}}}}
The key of an override can also include a version, or range of versions. To overridefoo
to1.0.0
, but only when it's a child ofbar@2.0.0
:
{"overrides":{"bar@2.0.0":{"foo":"1.0.0"}}}
You may not set an override for a package that you directly depend on unless both the dependency and the override itself share the exact same spec. To make this limitation easier to deal with, overrides may also be defined as a reference to a spec for a direct dependency by prefixing the name of the package you wish the version to match with a$
.
{"dependencies":{"foo":"^1.0.0"},"overrides":{// BAD, will throw an EOVERRIDE error// "foo": "^2.0.0"// GOOD, specs match so override is allowed// "foo": "^1.0.0"// BEST, the override is defined as a reference to the dependency"foo":"$foo",// the referenced package does not need to match the overridden one"bar":"$foo"}}
engines
You can specify the version of node that your stuff works on:
{"engines":{"node":">=0.10.3 <15"}}
And, like with dependencies, if you don't specify the version (or if you specify "*" as the version), then any version of node will do.
You can also use the "engines" field to specify which versions of npm are capable of properly installing your program. For example:
{"engines":{"npm":"~1.0.20"}}
Unless the user has set theengine-strict
config flag, this field is advisory only and will only produce warnings when your package is installed as a dependency.
os
You can specify which operating systems your module will run on:
{"os":["darwin","linux"]}
You can also block instead of allowing operating systems, just prepend the blocked os with a '!':
{"os":["!win32"]}
The host operating system is determined byprocess.platform
It is allowed to both block and allow an item, although there isn't any good reason to do this.
cpu
If your code only runs on certain cpu architectures, you can specify which ones.
{"cpu":["x64","ia32"]}
Like theos
option, you can also block architectures:
{"cpu":["!arm","!mips"]}
The host architecture is determined byprocess.arch
private
If you set"private": true
in your package.json, then npm will refuse to publish it.
This is a way to prevent accidental publication of private repositories. If you would like to ensure that a given package is only ever published to a specific registry (for example, an internal registry), then use thepublishConfig
dictionary described below to override theregistry
config param at publish-time.
publishConfig
This is a set of config values that will be used at publish-time. It's especially handy if you want to set the tag, registry or access, so that you can ensure that a given package is not tagged with "latest", published to the global public registry or that a scoped module is private by default.
Seeconfig
to see the list of config options that can be overridden.
workspaces
The optionalworkspaces
field is an array of file patterns that describes locations within the local file system that the install client should look up to find eachworkspace that needs to be symlinked to the top levelnode_modules
folder.
It can describe either the direct paths of the folders to be used as workspaces or it can define globs that will resolve to these same folders.
In the following example, all folders located inside the folder./packages
will be treated as workspaces as long as they have validpackage.json
files inside them:
{"name":"workspace-example","workspaces":["./packages/*"]}
Seeworkspaces
for more examples.
DEFAULT VALUES
npm will default some values based on package contents.
"scripts": {"start": "node server.js"}
If there is a
server.js
file in the root of your package, then npm will default thestart
command tonode server.js
."scripts":{"install": "node-gyp rebuild"}
If there is a
binding.gyp
file in the root of your package and you have not defined aninstall
orpreinstall
script, npm will default theinstall
command to compile using node-gyp."contributors": [...]
If there is an
AUTHORS
file in the root of your package, npm will treat each line as aName <email> (url)
format, where email and url are optional. Lines which start with a#
or are blank, will be ignored.