speech.rb - OpenAIのAPIを使ってテキストファイルを音声MP3に変換するRubyスクリプト

speech.rb

# ◆Ruby OpenAI
# https://github.com/alexrudall/ruby-openai
# $ gem install ruby-openai
#
# ◆Text to speech - OpenAI API (voice options)
# https://platform.openai.com/docs/guides/text-to-speech/voice-options
#
# ◆API keys
# https://platform.openai.com/api-keys
# $ grep OPENAI_API_KEY ~/.bash_profile
# export OPENAI_API_KEY='sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx'
#
# ◆Billing
# https://platform.openai.com/account/billing/overview

require "openai"

if ARGV.size != 3
  abort "Usage: ruby speech.rb voice input.txt output.mp3"
end

voice = ARGV[0]
input_txt = ARGV[1]
output_mp3 = ARGV[2]
input = File.read(input_txt)

# 引数の表示
puts "voice: #{voice}"
puts "input:\n#{input}"
puts "mp3: #{output_mp3}"

# クライアントの生成
client = OpenAI::Client.new(access_token: ENV["OPENAI_API_KEY"])

# 音声合成
response = client.audio.speech(
  parameters: {
    model: "tts-1",
    input: input,
    voice: voice,
  }
)

# 結果の書き込み
File.binwrite(output_mp3, response)