AWS DevOps Blog

Introducing Amazon CodeWhisperer for command line

The command line is used by over thirty million engineers to write, build, run, debug, and deploy software. However, despite how critical it is to the software development process, the command line is notoriously hard to use. Its output is terse, its interface is from the 1970s, and it offers no hints about the ‘right way’ to use it. With tens of thousands of command line applications (called command-line interfaces or CLIs), it’s almost impossible to remember the correct input syntax. The command line’s lack of input validation also means typos can cause unnecessary errors, security risks, and even production outages. It’s no wonder that most software engineers find the command line an error-prone and often frustrating experience.

Announcing Amazon CodeWhisperer for command line
Amazon CodeWhisperer for command line is a new set of capabilities and integrations for AI-powered productivity tool, Amazon CodeWhisperer, that makes software developers more productive in the command line. CodeWhisperer for command line modernizes the command line with features such personalized code completions, inline documentation, and AI natural-language-to-code translation. You don’t need to change the tools you use to start benefiting from CodeWhisperer for command line: it integrates directly with your existing tools, such as iTerm2 or the VS Code embedded terminal.

To get started, download CodeWhisperer for command line here (macOS only).

IDE-style completions for 500+ CLIs
CodeWhisperer for command line adds IDE-style completions for hundreds of popular CLIs like as Git, npm, Docker, MongoDB Atlas, and the AWS CLI. These typeahead completions increase your productivity by reducing the time spent typing repetitive or boilerplate commands. Inline documentation helps you understand CLI functionality without context-switching to the browser and interrupting your workflow.

Previously, typing a CLI command like git and hitting tab either wouldn’t show you any completions or would show an incomplete list of completions in a clunky interface without descriptions. Now, you can type git and see all the git subcommands, options, and arguments with descriptions, ordered by usage recency. You can also type cd to see a list of all your directories, npm install to see a list of all the node packages available to install, or aws to see a list of all the AWS CLI subcommands.

Demo of CodeWhisperer for command line CLI completion functionality showing a user getting autocompletions for cd and popular git subcommands, options, and arguments


Natural language-to-bash translation

CLI completions are great for tasks where you already know how to do something and just want to move faster. But what do you do when you’re trying to solve a problem and you’re not 100% sure how? Enter: cw ai!

The cw ai command lets you write a natural language instruction and CodeWhisperer will translate it to an instantly executable shell code snippet. For instance, imagine you want to copy a file from your local machine to Amazon Simple Storage Service (Amazon S3). You would write “copy all files in my current directory to s3” and CodeWhisperer will output aws s3 cp . s3://$BUCKET_NAME --recursive — now all you need to do is choose an S3 bucket. Natural language to bash translation is perfect for those workflows you occasionally have to do, but always forget the correct bash syntax like reversing a git commit, finding strings inside files with grep, or compressing files with tar. And just like with CLI completions, cw ai translator works great with the AWS CLI.

CodeWhisperer for command line demo showing a user typing the `cw ai` command, asking to copy all the files in their current directory to S3, and getting the correct command as output

Get started
CodeWhisperer for command line is available on macOS for all major shells (bash, zsh, and fish) and major terminal emulators such as Terminal, iTerm2, Hyper, and the built-in terminals in Visual Studio Code and JetBrains.

To get started, download CodeWhisperer for command line here. To learn more, visit the CodeWhisperer documentation