Skip to content

Sharing your environments

Flox provides three main ways of sharing environments with others:

  • Sharing environments with files:: Flox environments are shared via the .flox folder and often checked into version control.
  • Sharing environments on FloxHub: Flox environments are shared via the FloxHub and available to all command line commands (including RC files) with --remote username/environment.
  • Containers: Flox environments are containerized or used to create container images.

Sharing environments with files

New environments created with flox init will create a .flox folder in the directory where flox init was run. This folder contains everything required to run this environment on their system and can be sent to another user any way files are shared. It is most common to commit the .flox to a version controlled code repository such as git. If shared this way, the user needs only git clone the project and flox activate in the directory containing the .flox.

Note

If you are sharing an environment with a user on a different CPU architecture or OS from the person who created the environment, you may run into some issues where system-specific packages are not available on their OS. This can be fixed with a minor edit to the manifest, described in the manifest concept. If you get stumped, reach out to us on our forum for assistance.

Here is an example of sharing a project with files. The first person creates the environment:

$ mkdir example-project # (1)!
$ cd example-project
$ git init
...
$ flox init
...
  1. example-project is a stand-in for a git source code managed project.

Install packages:

$ flox install inetutils neovim curl
✅ 'inetutils' installed to environment example-project at /Users/youruser/example-project
✅ 'neovim' installed to environment example-project at /Users/youruser/example-project
✅ 'curl' installed to environment example-project at /Users/youruser/example-project

Add the .flox directory and commit the changes.

$ git add .flox
$ git commit -m "sharing flox environment"

Another developer on the same project can get started immediately with flox activate

$ git clone ..example-project
$ flox activate

Sharing environments on FloxHub

flox push for the first time

The flox push command makes it easy to share your environment using FloxHub. When you flox push for the first time, you can create an account on FloxHub for free and send your environment's manifest and metadata for easy sharing. 

$ flox push
✅  example-project successfully pushed to FloxHub

    Use 'flox pull youruser/example-project' to get this environment in any other location.
You can also view your new environment in FloxHub's web application.

Directly activating a remote environment

As the recipient, you can use the environment in a variety of ways depending on your needs. If you trust the user sending the environment, flox activate -r username/environment the environment directly. The first time you do this you will be offered a choice about trusting this user in the future.

$ flox activate -r youruser/example-project
Environment youruser/example-project is not trusted.

    flox environments do not run in a sandbox.
    Activation hooks can run arbitrary code on your machine.
    Environments need to be trusted to be activated.
? Do you trust youruser/example-project?
  Do not trust, ask again next time
  Do not trust, save choice
  Trust, ask again next time
  > Trust, save choice
  Show the manifest

Trusted environment youruser/example-project 

flox [youruser/example-project] $ telnet --version
telnet (GNU inetutils) 2.5
...

Pulling a remote environment and pushing updates

If you intend to use the same environments across multiple projects or you want to stage a change to the remote environment, you may want to flox pull it instead.

flox pull adds a .flox folder to the directory you are in that is linked to the remote environment. When using a FloxHub environment in multiple projects it allows centralized management of the dependencies used across these projects--future updates to the environment will get automatically picked up by projects using the environment:

$ cd similar-example-project
$ flox pull youruser/example-project
✨  Pulled youruser/example-project from https://hub.flox.dev

    You can activate this environment with 'flox activate'

It can also be useful to flox pull when you need to work on a new version of the environment without editing what everyone is using live. After pulling an environment you can install changes to it and, when you're ready, flox push them for everyone using the environment if the environment is unlocked:

$ flox install yarn
✅ 'yarn' installed to environment youruser/example-project at /Users/youruser/similar-example-project
$ flox push
✅  Updates to example-project successfully pushed to FloxHub

    Use 'flox pull youruser/example-project' to get this environment in any other location.

Note

Right now, only environment owners can push edits to their environments.

Always using the same environment across multiple devices

It can be useful to share the same environment across multiple machines where an install to one will install to the others. To do this, you need to flox push your environment and add a flox activate -r to your terminal's RC file. Let's look at an example using the environment youruser/example-project for a zsh user, so we can have everything we installed automatically on multiple machines.

Edit your rc file using an editor of choice.

$ vim ~/.zshrc

Append this line to your shell's rc file or fish.config at the bottom.

For bash .bashrc or zsh .zshrc
eval "$(flox activate -r youruser/example-project)"
For fish config.fish
eval (flox activate -r youruser/example-project) | source

Don't forget to open a new terminal window or, for zsh, reload your RC file. 

$ source ~/.zshrc

Now all new windows will open into your FloxHub environment. As the environment changes, the latest will be activated.

Sharing with containers

Flox can render that environment as an OCI container runtime suitable for use with containerd, docker, kubernetes, nomad, and more.

flox containerize is currently experimental and only supported on Linux

Let's create a container image from the example-environment we have been working with.

To render your environment to a container, run flox containerize. This command will pipe the container image to stdout for usage with docker load:

$ flox containerize | docker load # (1)!
...
Building container...
Done.
No 'fromImage' provided
Creating layer 1 from paths: [...]
...
Adding manifests...
Done.
✨  Container written to '/home/youruser/default-container.tar.gz'
  1. flox containerize will stream generated docker image to docker load command.

Why so many layers?

By default flox splits every dependency into a different layers, which allows better dependency sharing and faster iteration.

Now let's run a command from our image:

$  docker run --rm -it flox-env-container -- telnet --version
telnet (GNU inetutils) 2.5
...

Where to next?