clx is a useful utility that generates CLI commands using AI models for common operational tasks. It's inspired by Shayan's clx and enhanced to support configurable models and multiple backends including Groq and OpenAI.

Download the latest binary from Releases.

go install github.com/mr-karan/clx@latest

➜ clx show disk usage of current directory
Showing disk usage of the current directory
$ du -sh .

➜ clx kill process on port 8000
Kill process running on port 8000.
$ lsof -t -i :8000 | xargs kill -9

➜ clx backup a directory
Backup a directory to a tarball archive
$ tar -czf backup.tar.gz /path/to/directory

➜ clx
NAME:
   clx - a CLI code generator

USAGE:
   clx [global options] command [command options]

COMMANDS:
   help, h  Shows a list of commands or help for one command

GLOBAL OPTIONS:
   --model value, -m value    Model to use for generating responses (default: "gpt-4-turbo")
   --backend value            Backend service to use (openai, groq) (default: "openai")
   --timeout value, -t value  Timeout for API requests (default: 30s)
   --config value, -c value   Path to config file
   --help, -h                 show help

clx can be configured using a config.toml file, which allows users to set default values for the backend service, model, and request timeout. Here’s an example of what the config.toml might look like:

backend = "groq" # Backend service (openai, groq)
model = "llama3-70b-8192" # Model to use
timeout = "30s" # Timeout for API requests

This file is read at runtime, and its settings are applied unless overridden by command-line flags. If the file does not exist, clx will create a default one using built-in settings. The file is located at $HOME/clx.toml.