Me and ChatGPT created a Bash script for text-to-speech using OpenAI TTS

Hi Dev.to community!

ChatGPT became an essential helper in writing and analyzing a code. But this specific case makes ChatGPT incredibly helpful!

If you caught yourself crunching the Bash syntax to make a handy automation on your Linux or Mac, you should hire ChatGPT.

Just look at the output below! I asked GPT to create a text-to-speech Bash script that works similar to say command on Linux. It took just a minute to build this useful, flexible and beautiful CLI tool out of an idea.

	#!/bin/bash
	# Function to display help
	usage() {
	echo "Usage: $0 [-o output_file] [-v voice] [-m model] [-i input_file] \"input_string\""
	exit 1
	}
	# Default values
	output_file="speech.mp3"
	voice="echo"
	model="tts-1"
	input_file=""
	input_string=""
	# Parse command-line options
	while getopts 'o:v:m:i:h' flag; do
	case "${flag}" in
	o) output_file="${OPTARG}" ;;
	v) voice="${OPTARG}" ;;
	m) model="${OPTARG}" ;;
	i) input_file="${OPTARG}" ;;
	h) usage ;;
	*) usage ;;
	esac
	done
	# Check for remaining arguments
	if [ -z "$input_file" ]; then
	if [ $OPTIND -gt $# ]; then
	echo "Error: Missing input string"
	usage
	else
	input_string="${@:$OPTIND:1}"
	fi
	else
	if [ -f "$input_file" ]; then
	input_string=$(<"$input_file")
	else
	echo "Error: Input file does not exist"
	exit 1
	fi
	fi
	# Prepare parameters for OpenAI API call
	PARAM=$(jq -n -c --arg model "$model" --arg voice "$voice" --arg input "$input_string" '$ARGS.named')
	# Call OpenAI API to generate speech
	curl https://api.openai.com/v1/audio/speech \
	-H "Authorization: Bearer $OPENAI_API_KEY" \
	-H "Content-Type: application/json" \
	-d "$PARAM" \
	--output "$output_file"

view raw say.sh hosted with ❤ by GitHub

Quick tip: How to create a Bash/Python/etc. script using ChatGPT

1. Start by asking to create a basic script. Do not provide details on what the script will do. Just tell more about script interface: CLI arguments, parameters, environmental variables etc.
2. Tell ChatGPT apply necessary modifications and improvements to make this script do the actual workload. Provide sample code and ask to integrate some code pieces found over the web if needed.
3. Well done!

Read the full story in Medium.