NVS is a cross-platform utility for switching between different versions and forks of Node.js. NVS is itself written in node JavaScript.

This tool is obviously inspired by other node version manager tools, especially nvm, from which it borrows a lot of ideas and some command-line syntax.

Following are basic setup instructions. For more details and options for setting up NVS, refer to the Setup page.

A Windows Installer (MSI) package is available from the NVS releases page on GitHub.

You can use winget to install it (available by default in Windows 11):

winget install jasongin.nvs

You can also use chocolatey to install it:

choco install nvs

Specify the installation path, clone the repo, and source the install command:

export NVS_HOME="$HOME/.nvs"
git clone https://github.com/jasongin/nvs "$NVS_HOME"
. "$NVS_HOME/nvs.sh" install

The nvs.sh script adds an nvs shell function to the environment. Afterward the tool should be invoked as just nvs without any path. The install command adds lines to your ~/.bashrc, ~/.profile, or ~/.zshrc file to source nvs.sh, so that the nvs function is available in future shells.

For ksh, the source nvs.sh needs to be in your ~/.kshrc or wherever $ENV points.

NVS can be used in a CI environment such as AppVeyor or Travis CI, to test a Node.js app or library on any version of Node.js that NVS can install.

To add the latest version of node:

$ nvs add latest

Or to add the latest LTS version of node:

$ nvs add lts

Then run the nvs use command to add a version of node to your PATH for the current shell:

$ nvs use lts
PATH += ~/.nvs/node/6.9.1/x64

To add it to PATH permanently, use nvs link:

$ nvs link lts

A version or filter consists of a complete or partial semantic version number or version label ("lts", "latest", "Argon", etc.), optionally preceded by a remote name, optionally followed by an architecture, separated by slashes. Examples: "lts", "4.6.0", "6/x86", "node/6.7/x64".

Refer to the docs for more details about each command.

When invoked with no parameters, nvs displays an interactive menu for switching and downloading node versions.

NVS uses console-menu, a module originally written for this project then published separately.

Visual Studio Code can use NVS to select a node version to use when launching or debugging. In launch.json (in the folder .vscode located on the project's root folder), add a "runtimeArgs" attribute with an NVS version string and a "runtimeExecutable" attribute that refers to nvs.cmd (Windows) or nvs (Mac, Linux). (You may need to specify an absolute path such as "${env:HOME}/.nvs/nvs" if NVS is not in VS Code's PATH.)

Example: Configure launch.json so VS Code uses NVS to launch node version 6.10:

  "configurations": [
    {
      "type": "node",
      "request": "launch",
      "name": "Launch Program",
      "program": "${file}",
      "args": [ ],
      "runtimeArgs": [ "6.10" ],
      "windows": { "runtimeExecutable": "nvs.cmd" },
      "osx": { "runtimeExecutable": "nvs" },
      "linux": { "runtimeExecutable": "nvs" }
    },
  ]

Or, remove the version string from "runtimeArgs" to get the version from a .node-version file in the project directory. For more details, see the NVS VS Code documentation or run nvs help vscode.

The nvs remote command allows configuration of multiple named download locations. NVS manages versions from different remote locations separately, so there is no risk of version collisions. By default there is a single remote pointing to Node.js official releases:

$ nvs remote
default  node
node     https://nodejs.org/dist/

This makes it possible to get builds from other sources. The following command sequence adds a remote entry for nightly builds, lists nightly builds, and adds a build:

$ nvs remote add nightly https://nodejs.org/download/nightly/
$ nvs lsr nightly/13
nightly/13.1.1-nightly20191120c7c566023f
...
$ nvs add nightly/13

Other remote sources are supported, for example:

nvs remote add iojs https://iojs.org/dist/
nvs remote add chakracore https://nodejs.org/download/chakracore-release/

An alias refers to a combination of a remote name and a semantic version. (Processor architectures are not aliased.) When setting an alias, the remote name may be omitted, in which case the alias refers to the default remote. An alias may be used in place of a version string in any of the other commands.

$ nvs alias myalias 6.7.0
$ nvs alias
myalias default/6.7.0
$ nvs run myalias --version
v6.7.0
$ nvs which myalias
~/.nvs/node/6.7.0/x64/bin/node
$ nvs which myalias/32
~/.nvs/node/6.7.0/x86/bin/node

An alias may also refer to a local directory, enabling NVS to switch to a local private build of node.

In either Bash or PowerShell, NVS can automatically switch the node version in the current shell as you change directories. This function is disabled by default; to enable it run nvs auto on. Afterward, whenever you cd or pushd under a directory containing a .node-version or an .nvmrc file then NVS will automatically switch the node version accordingly, downloading a new version if necessary. When you cd out to a directory with no .node-version or .nvmrc file anywhere above it, then the default (linked) version is restored, if any.

~$ nvs link 6.9.1
~/.nvs/default -> ~/.nvs/node/6.9.1/x64
~$ nvs use
PATH += ~/.nvs/default/bin
~$ nvs auto on
~$ cd myproject
PATH -= ~/.nvs/default/bin
PATH += ~/.nvs/node/4.6.1/x64/bin
~/myproject$ cd ..
PATH -= ~/.nvs/node/4.6.1/x64/bin
PATH += ~/.nvs/default/bin

This feature is not available in Windows Command Prompt. Use PowerShell instead.

If your shell isn't compatible with automatic switching or you'd prefer to switch manually but still take advantage of any .node-version or .nvmrc files, you can run nvs use with the version auto or just run nvs auto.

$ nvs use auto

is equivalent to

$ nvs auto

NVS uses a small amount of platform-specific shell code that bootstraps the tool by automatically downloading a private copy of node. The bootstrap code is just a few dozen lines each of Windows command script, Windows powershell script, and POSIX shell script. Besides bootstrapping, the shell scripts are also used to export PATH changes to the calling shell (which a separate node process cannot do). But all the code for querying available versions, downloading and installing node and matching npm, switching versions/architectures/engines, uninstalling, parsing and updating PATH, and more can be written in JavaScript, and mostly in a cross-platform way.

NVS downloads node builds under the directory specified by the NVS_HOME environment variable, or under the NVS tool directory if NVS_HOME is not set. Each build goes in a subdirectory based on the remote name, semantic version, and architecture, for example node/6.7.0/x64.

When you nvs use a version, the PATH of the current shell is updated to include that version's bin directory.

When using npm install -g or npm link with NVS-installed node, global modules are installed or linked into a version-specific directory. (NVS clears any NPM_CONFIG_PREFIX environment variable that may have been set.) This means when NVS switches versions it is also switching the set of available global modules. The nvs migrate command can migrate those global modules from one node version to another.

The nvs link command creates a symbolic directory link at $NVS_HOME/default that points to the specified version (or the current version from PATH at the time of the command). This can be useful when there is a need to configure a fixed path elsewhere.

On non-Windows platforms, a new shell that sources the nvs.sh script also sets PATH to include the default version, if a link is present. On Windows, the PATH environment variable is updated in the user profile, so that new shells will use the default version.

The nvs ls command lists all local node versions, marks the version currently in the path with a >, and marks the default (linked) version, if any, with a #. These may be the same or different. For example:

  node/4.5.0/x64
 #node/4.6.0/x64
 >node/6.7.0/x64

If $NVS_HOME is under a system path such as /usr/local or %ProgramFiles%, then the nvs link command additionally links into well-known Node.js system locations. (This is only allowed if there is not already a system-installed node.)

• On non-Windows platforms, symbolic links are created in /usr/local/bin for node, npm, and any globally-installed node modules that have executables. Note after installing or uninstalling global modules that include executables it may be necessary to run nvs link again to update the global links. Using NVS to link a different version of node (with different global modules) updates all the links accordingly.

• On Windows, a symbolic directory link is created at %ProgramFiles%\Nodejs, and that directory is added to the system PATH.

This system linking functionality is skipped when $NVS_HOME points to a non-system directory, because it would be wrong to create symlinks in system directories to user files.

NVS has no external dependencies beyond the private copy of node that it automatically downloads. Runtime JS package dependencies are minimal and are checked in to the repo to avoid the need for an npm install at bootstrapping time.