The commands are presented here in alphabetic order, but the most important ones are init, shell](#shell), and [run.


The command avatar export-env, if executed inside an Avatar-CLI project directory, will output a list of shell export statements that define an "environment".

It can be used in scripts to activate the tools managed by Avatar-CLI.

If you are using Bash scripts, you could "activate" the environment this way:


source <(avatar export-env)

If you want full compatibility with POSIX shells, then you have to first create a temporary file and then source it:


avatar export-env > /your/temporary/file
. /your/temporary/file # Notice how we use '.' and not 'source'


As you might expect, typing avatar help in the command line will show you some help information:

avatar 0.18.1

    avatar <SUBCOMMAND>

    -h, --help       Prints help information
    -V, --version    Prints version information

    export-env    Prints shell variable exports to create a new Avatar-CLI session. Useful for scripts.
    help          Prints this message or the help of the given subcommand(s)
    init          It generates a new Avatar-CLI project configuration
    install       It 'installs' all the project stated dependencies
    run           Executes a wrapped project tool without having to enter into a subshell
    shell         Starts a new subshell exposing the wrapped project tools

You can also type avatar help [subcommand] to get more detailed information, as an example, for avatar help run you would get:

Executes a wrapped project tool without having to enter into a subshell

    avatar run <program_name> [program_args]...

    -h, --help       Prints help information
    -V, --version    Prints version information



In case you are in a directory that has not been already initialized as an Avatar-CLI project, by typing avatar init the project will be initialized.

The initialization consists on creating a new directory .avatar-cli and placing inside a new YAML file called Avatarfile. That's the file that we'll modify to tweak our project.


The command avatar install will update the "lock files" and ensure that all the required OCI/Docker images are available.

In general terms, you won't have to use this command almost ever, as it's implicitly triggered by the subcommands run and shell if needed.


Assuming that you have an already initialized project where you defined the the software you want to run in your Avatarfile, you could execute the desired program with the following command:

avatar run program_name

In case you want to pass parameters to the containerized application, you can do it like this (notice the -- between the program name and its parameters):

avatar run program_name -- param_1 param_2 ... param_n

As you can see, running the containerized applications like this can feel a little bit cumbersome. The commands are slightly longer and complicated than when we are running non-containerized applications... and that's why we have the subcommand shell to overcome this issue.


When you type avatar shell, a new subshell will be created, where certain environment variables will be available, and the containerized cli tools that you configured in the Avatarfile will be available to you as if they were installed in your system.

For example, assuming that we don't have NodeJS installed in our system but we configured it in our project, then it will be available to us:

» node
node: command not found

# With this command, we go into a new subshell...
» avatar shell

# And now `node` is available to us :D
» node --version

» echo "console.log('hello')" | node