CLI Scripting
The core runtime allows you to build CLI scripting applications

BoxLang is a modern, dynamic scripting language built for more than just simple automation—it empowers you to create full-fledged, high-performance CLI applications with ease. Designed to run seamlessly on the JVM, BoxLang provides powerful scripting capabilities, a rich standard library, and first-class support for modular development.
Whether you're automating repetitive tasks, building interactive command-line tools, or developing complex CLI-driven workflows, BoxLang offers the flexibility, expressiveness, and performance you need. With intuitive syntax, robust error handling, and seamless integration with Java and other JVM-based technologies, BoxLang makes CLI scripting more efficient and enjoyable.
BoxLang CLI Entry Points & Conventions
BoxLang supports multiple ways to execute code from the command line, making it a flexible tool for scripting, automation, and app development. Here’s a summary of the main entry points and conventions:
File Execution
You can execute any supported file type directly:
*.bx
— BoxLang class with amain()
method*.bxs
— BoxLang script*.bxm
— BoxLang template*.cfs
/*.cfm
— CFML script/template (requiresbx-compat-cfml
module)*.sh
— Shebang script (with#!/usr/bin/env boxlang
)
BoxLang will automatically detect and run the correct entry point, including shebang scripts and classes with a main()
method.
Script Files
With BoxLang, you can execute a few types of files right from any OS CLI by adding them as the second argument to our boxlang
binary:
*.bx
BoxLang classes with a main()
method
*.bxs
BoxLang scripts
*.bxm
BoxLang Templating Language
*.cfs
CFML scripts (If using the bx-compat-cfml
module)
*.cfm
CFML templates (If using the bx-compat-cfml
module)
*.sh
Shebang scripts using boxlang
as the environment
Please note that you will need the bx-compat-cfml
module if you want to execute CFML scripts
Here are some examples of executing the files. Just pass in the file by relative or absolute path location.
boxlang task.bx
boxlang myscript.bxs
boxlang mytemplate.bxm
boxlang /full/path/to/test.bxs
boxlang /full/path/to/Task.bx
Other Scopes
Please note that you have access to other persistent scopes when building CLI applications:
application
- This scope lives as long as your application lives as well, but it is technically attached to anApplication.bx
file that activates framework capabilities for your application.request
- A scope that matches a specific request for your application. We also get one per CLI app since there is no concept of sessions or user state. There is always only one request. It would be up to you to create a session-like mechanism if you need to persist state across multiple executions.server
- A scope that lives as long as the CLI app is running. This is useful for storing global state or configuration that should persist across multiple requests or executions.
For CLI applications, we recommend you use the server
or request
scope for singleton persistence. Also note that you can use all the caches as well for persistence. You can use application
scope if you have an Application.bx.
Executing Classes
BoxLang allows you to execute any *.bx
class as long as it has a method called main()
by convention. All the arguments passed into the file execution will be collected and passed into the function via the args
argument.
class {
function main( args = [ ] ) {
println( "Hola from my task! #now()#" );
println( "The passed args are: " );
println( args );
}
}
The args
argument is an array and it will contain all the arguments passed to the execution of the class.
boxlang task.bx hola --many options=test
If you execute this function above, the output will be:
Hola from my task! { ts '2025-02-11 22:15:44' }
The passed args are:
[
hola,
--many,
options=test
]
Class executions are a great way to build tasks that have a deterministic approach to execution. We parse the arguments for you, and you can focus on building your task.
Executing Scripts / Templates
In addition to executing classes, you can execute *.bxs
scripts that can do your bidding. The difference is that this is a flat source code script that executes from the top down. It can contain functions, scope usage, imports, and create any class.
message = "Hola from my task! #now()#";
println( message );
println( "The passed args are: " );
println( CLIGetArgs( ) );
Then, if we execute it, we can see this output:
╰─ boxlang hello.bxs hola luis=majano --test
Hola from my task! { ts '2025-02-11 22:29:44' }
The passed args are:
{
positionals : [
hola,
luis=majano
],
options : {
test : true
}
}
What do you see that's different? We don't have the incoming arguments as an argument since it's a script. However, we can use the CLIGetArgs()
BIF, and it will give you a structure of two keys:
positionals
- An array of positional values passed to the scriptoptions
- Name value pairs detected as options
message = "Hola from my task! #now()#";
println( message );
println( "The passed args are: " );
println( server.cli.parsed );
Here is the output:
╰─ boxlang hello.bxs hola luis=majano --test
Hola from my task! { ts '2025-02-11 22:29:44' }
The passed args are:
{
positionals : [
hola,
luis=majano
],
options : {
test : true
}
}
Please note that executing templates is the same as scripts, but your template uses templating language instead, which can be helpful if you produce some markup (HTML, Markdown, etc.)
SheBang Scripts
SheBang scripts are text files containing a sequence of commands for a computer operating system. The term "shebang" refers to the #!
characters at the beginning of the script, which specify the interpreter that should be used to execute the script. These scripts are commonly used in Unix-like operating systems to automate tasks. You can run scripts directly from the command line using a shebang line without explicitly invoking the interpreter. BoxLang supports these scripts, so the OS sees them as just pure shell scripts, but you are coding in BoxLang scripting.
A SheBang script is just basically a *.bxs
script.
#!/usr/bin/env boxlang
println( "Hello World! #now()#" );
println( CLIGetArgs( ) );
As you can see from the sample above, the first line is what makes it a SheBang script the operating system can use. It passes it to the boxlang
binary for interpretation. Also, note that you can pass arguments to these scripts like any other script and the CLIGetArgs()
or the server.cli.parsed
variables will be there for you to use.
# Execute the script
./hola.sh
# Execute it with a name argument and a simple option
./hola.sh --name=luis -d
BoxLang CLI Options and Flags
BoxLang provides a comprehensive set of CLI options and flags for various development and execution scenarios. All BoxLang-specific options are prefixed with --bx-
to avoid conflicts with other tools.
Global Options
-h, --help
Show help message and exit
--version
Show version information and exit
--bx-debug
Enable debug mode with timing information
--bx-config <PATH>
Use custom BoxLang configuration file
--bx-home <PATH>
Set BoxLang runtime home directory
--bx-code <CODE>
Execute inline BoxLang code directly
--bx-printAST
Print Abstract Syntax Tree for code analysis
--bx-transpile
Transpile BoxLang code to Java
Environment Variables
You can also control BoxLang behavior using environment variables:
BOXLANG_DEBUG=true
Enable debug mode
BOXLANG_CONFIG=/path/config.json
Override configuration file path
BOXLANG_HOME=/path/to/home
Set runtime home directory
BOXLANG_TRANSPILE=true
Enable transpile mode
BOXLANG_PRINTAST=true
Enable AST printing
Examples of CLI Options
# Execute with debug mode enabled
boxlang --bx-debug myapp.bx
# Use custom configuration file
boxlang --bx-config ./custom.json myapp.bx
# Execute inline code
boxlang --bx-code "println( 'Hello BoxLang!' )"
# Print AST for code analysis
boxlang --bx-printAST --bx-code "x = 1 + 2"
# Combined options
boxlang --bx-debug --bx-config ./custom.json myapp.bx
Action Commands
BoxLang includes several powerful action commands for development workflows:
Compile Command
Pre-compile BoxLang templates to class files for improved performance:
# Get help for compile command
boxlang compile --help
# Compile source directory to target
boxlang compile --source ./src --target ./compiled
CF Transpile Command
Transpile ColdFusion code to BoxLang syntax:
# Get help for cftranspile command
boxlang cftranspile --help
# Transpile legacy CF code to BoxLang
boxlang cftranspile --source ./legacy --target ./modern
Feature Audit Command
Audit your code for BoxLang feature compatibility:
# Get help for featureaudit command
boxlang featureaudit --help
# Audit code and generate report
boxlang featureaudit --source ./myapp --output report.json
Runtime Mode Detection
BoxLang provides several ways to detect the runtime execution context:
Server Scope Information
The server
scope contains detailed information about the runtime environment:
// Check if running in CLI mode
if ( server.boxlang.cliMode ) {
println( "Running in CLI mode" );
}
// Check if running from JAR
if ( server.boxlang.jarMode ) {
println( "Running in JAR mode" );
}
// Get runtime home directory
println( "Runtime home: " & server.boxlang.runtimeHome );
CLI-Specific Information
When running in CLI mode, additional CLI-specific information is available:
// Access CLI execution details
println( "Execution path: " & server.cli.executionPath );
println( "Command: " & server.cli.command );
println( "Raw args: " & server.cli.args.toString( ) );
println( "Parsed args: " & server.cli.parsed.toString( ) );
Inline Code Execution
You can execute BoxLang code directly from the CLI using the --bx-code
flag:
boxlang --bx-code "println( 'Hello from BoxLang!' )"
Scheduler Files
You can run BoxLang scheduler files using the schedule
action command. The file must be a .bx
component with scheduler definitions. The scheduler will run continuously until you press Ctrl+C
.
boxlang schedule ./schedulers/MainScheduler.bx
REPL Mode
When no arguments are provided, BoxLang starts in REPL mode (Read-Eval-Print-Loop):
Interactive environment for testing and development
Type expressions and see results immediately
Supports multi-line expressions and complex code
Press
Ctrl+C
to exit REPL mode
# Start REPL mode
boxlang
# You'll see the REPL prompt
BoxLang> println( "Hello from REPL!" )
Hello from REPL!
BoxLang>
CLI Built-In Functions
BoxLang also gives you several built-in functions for interacting with the CLI:
CLIClear( ):void
- Clears the consoleCLIGetArgs( ):struct
- Return a structure of the parsed incoming argumentsCLIRead( [ prompt ] ):any
- Read input from the CLI and return the valueCLIExit( [ exitCode=0 ] )
- Do aSystem.exit( )
with the passed-in exit code
Please note that you have a wealth of built-in functions and components that you can use to build your scripts.
Parsed Arguments
BoxLang automatically parses incoming arguments into a structured format when using the CLIGetArgs( )
BIF or by accessing the server.cli.parsed
variable.
The parsed structure contains:
options
- A structure of the options (name-value pairs) used to invoke the scriptpositionals
- An array of the positional arguments used to invoke the script
CLI Argument Formats
BoxLang supports standard CLI argument formats:
--option
Boolean option set to true
--debug
--option=value
Option with a value
--config=myfile.json
--option="value"
Option with quoted value
--message="Hello World"
--option='value'
Option with single quoted value
--message='Hello World'
-o=value
Shorthand option with value
-c=config.json
-o
Shorthand boolean option set to true
-v
--!option
Negation option set to false
--!verbose
--no-{option}
Negation option set to false
--no-debug
Multi-Character Shorthand Options
You can combine multiple single-character options:
# This creates: a=true, b=true, c=true
boxlang myscript.bxs -abc
Parsing Examples
For the following CLI command:
boxlang myscript.bxs --debug --!verbose --config=prod.json -o='/path/to/file' -v my/path/template
The parsed structure will be:
{
"options": {
"debug": true,
"verbose": false,
"config": "prod.json",
"o": "/path/to/file",
"v": true
},
"positionals": [ "my/path/template" ]
}
Accessing Parsed Arguments
// In a script file (.bxs)
var cliArgs = CLIGetArgs( );
// or
var cliArgs = server.cli.parsed;
// Check for options
if ( cliArgs.options.debug ) {
println( "Debug mode enabled" );
}
// Process positional arguments
cliArgs.positionals.each( function( arg ) {
println( "Processing: " & arg );
} );
Advanced CLI Information
The server.cli
structure provides comprehensive CLI context:
// Full CLI context information
var cliInfo = server.cli;
println( "Execution Path: " & cliInfo.executionPath );
println( "Full Command: " & cliInfo.command );
println( "Raw Arguments: " & cliInfo.args.toString( ) );
println( "Parsed Options: " & cliInfo.parsed.options.toString( ) );
println( "Positional Args: " & cliInfo.parsed.positionals.toString( ) );
Ground Rules for CLI Arguments
Options are prefixed with
--
(long form) or-
(short form)Shorthand options can be combined (e.g.,
-abc
=-a -b -c
)Options can be negated with
--!
or--no-
prefixValues can be assigned with
=
and optionally quotedRepeated options will override previous values
Everything after options are treated as positional arguments
Reading Input
You can easily read input from users by using our handy CLIRead( )
bif. You can also pass in a prompt
as part of the method call.
var exit = cliRead( "Do you want to continue? (Y/N)" ).trueFalseFormat( );
if ( exit ) {
cliExit( );
}
Producing Output
As you navigate all the built-in functions and capabilities of BoxLang, let's learn how to produce output to the system console.
printLn( )
- Print with a line break to System outprint( )
- Print with no line break to System outwriteOutput( ), echo( )
- Writes to the output buffer (Each runtime decides what its buffer is. The CLI is the system output, the Web is the HTML response buffer, etc)writeDump( )
- Takes any incoming output and will serialize to a nice string output representation. This will also do complex objects deeply.
println( "Time is #now()#" );
I get the output:
╰─ boxlang test.bxs
Time is { ts '2024-05-22 22:09:56' }
Hooray! You have executed your first script using BoxLang. Now let's build a class with a main( args=[ ] )
convention. This is similar to Java or Groovy.
class {
function main( args=[ ] ) {
println( "Task called with " & arguments.toString( ) );
writedump( args );
}
}
You can now call it with zero or more arguments!
╰─ boxlang Task.bx
Task called with { ARGS=[ ] }
╰─ boxlang Task.bx boxlang rocks
Task called with { ARGS=[ boxlang, rocks ] }
Piping code
You can also pipe statements into the BoxLang binary for execution as well. This assumes script, not tags.
echo "2+2" | java -jar boxlang-1.0.0.jar
echo "2+2" | boxlang
or
# on *nix
cat test.cfs | java -jar boxlang-1.0.0.jar
cat test.cfs | boxlang
# on Windows
type test.cfs | java -jar boxlang-1.0.0.jar
type test.cfs | boxlang.bat
Module CLI Apps
BoxLang allows you to build CLI applications as modules, making it easy to package, share, and execute reusable command-line tools. To create a module CLI app, simply add a main( args )
method to your module's ModuleConfig.bx
file.
When you want to execute a module as a CLI app, use the following convention:
module:{name}
- This will execute the module'sModuleConfig.main( args )
method, passing any CLI arguments to it.
For example, if you have a module named mytools
, you can run its CLI entry point like this:
boxlang module:mytools arg1 --option=value
This will invoke the main( args )
method in ModuleConfig.bx
of the mytools
module, with all CLI arguments available in the args
array.
Example: ModuleConfig.bx
class {
function main( args = [ ] ) {
println( "Module CLI called with args:" );
writedump( args );
// Your CLI logic here
}
}
This approach lets you build modular CLI utilities that can be distributed and executed just like standalone scripts or classes. You can leverage all BoxLang features, scopes, and built-in functions inside your module CLI apps.
Embedding Modules in a CLI App
BoxLang also allows you to embed modules inside your CLI application for distribution and local usage. This is different from creating a CLI app that executes a module's main()
method. Embedding modules means your CLI app can include and use additional BoxLang modules as dependencies, making your CLI tool more powerful and modular.
To embed modules, use the boxlang_modules
folder convention in your CLI app directory. You can install modules locally into this folder using the install-bx-module
installer script with the --local
flag:
install-bx-module bx-pdf bx-image --local
When your CLI app runs, BoxLang will check the boxlang_modules
folder first for available modules, then fall back to the OS home modules. This allows you to package all required modules with your CLI app for easy distribution and predictable behavior.
Example directory structure:
mycliapp/
myscript.bxs
boxlang_modules/
bx-pdf/
bx-image/
Your CLI scripts and classes can then use any embedded modules as if they were installed globally.
Additional Resources and Examples
Modern Development Workflow
BoxLang's CLI capabilities make it ideal for modern development workflows:
// example-workflow.bx
class {
function main( args = [ ] ) {
var cliArgs = CLIGetArgs( );
// Environment detection
var isDev = cliArgs.options.env == "development";
var isDebug = cliArgs.options.debug ?: false;
if ( isDebug ) {
println( "🐛 Debug mode enabled" );
println( "⚙️ Runtime Info:" );
println( " - CLI Mode: " & server.boxlang.cliMode );
println( " - JAR Mode: " & server.boxlang.jarMode );
println( " - Runtime Home: " & server.boxlang.runtimeHome );
}
// Process based on environment
if ( isDev ) {
runDevelopmentTasks( );
} else {
runProductionTasks( );
}
}
private function runDevelopmentTasks( ) {
println( "🔨 Running development tasks..." );
// Development-specific logic
}
private function runProductionTasks( ) {
println( "🚀 Running production tasks..." );
// Production-specific logic
}
}
Usage:
# Development mode with debug
boxlang --bx-debug example-workflow.bx --env=development --debug
# Production mode
boxlang example-workflow.bx --env=production
Integration Examples
BoxLang CLI can integrate with various tools and workflows:
# CI/CD Pipeline Integration
boxlang --bx-code "
println( 'Starting CI/CD Pipeline...' );
var result = runTests( );
if ( !result.success ) {
cliExit( 1 );
}
deployApplication( );
println( 'Pipeline completed successfully!' );
"
# Database Migrations
boxlang migrate.bx --action=up --env=production
# Code Generation
boxlang generate.bx --type=component --name=UserService
# Environment Setup
boxlang setup.bx --install-deps --configure-db
Performance and Debug Information
When using --bx-debug
, BoxLang provides detailed performance metrics:
boxlang --bx-debug myapp.bx
This outputs:
Execution timing information
Memory usage statistics
Runtime initialization details
AST parsing time
Module loading performance
Community Resources
📚 Documentation: https://boxlang.ortusbooks.com/
💬 Community Forum: https://community.ortussolutions.com/c/boxlang/42
💾 GitHub Repository: https://github.com/ortus-boxlang
🌐 Official Website: https://boxlang.io
🎯 Examples Repository: https://github.com/ortus-boxlang/bx-demos
Dad Joke Script
Thanks to our evangelist Raymond Camden, we have a cool dad joke script you can find in our demos: https://github.com/ortus-boxlang/bx-demos
class {
variables.apiURL = "https://icanhazdadjoke.com/";
/**
* The first argument is a term to search dad jokes on, if not provided, a random dad joke will be fetched.
* Example: boxlang DadJoke.bx dad
* Example: boxlang DadJoke.bx
*/
function main( args = [ ] ) {
// Use elvis operator to check if a term was passed, else, use an empty string
var term = ( args[ 1 ] ?: "" ).trim( );
if ( !term.isEmpty( ) ) {
apiURL &= "search?term=" & term.urlEncodedFormat( );
}
println( "Getting dad joke for term [#term#], please wait..." );
bx:http url=apiURL result="result" {
bx:httpparam type="header" name="Accept" value="application/json";
}
var data = JSONDeserialize( result.fileContent );
// possible none were found, use safe navigation operator
if ( data?.results?.len( ) == 0 ) {
println( "No jokes found for term: #term#" );
return cliExit( );
}
// If we searched for a term, we need to get a random joke from the results, otherwise, just .joke
var joke = term.isEmpty( ) ? data.joke : data.results[ randRange( 1, data.results.len( ) ) ].joke;
println( joke );
}
}
Now you execute it
// Random joke
boxlang DadJoke.bx
// Term jokes
boxlang DadJoke.bx ice
Let's modify it now so that we can prompt the user for the term using the CLIRead( )
BIF instead of passing it:
var term = ( CLIRead( "What search term would you like to use? (Leave blank for random joke)" ) ).trim( );
Last updated
Was this helpful?