rb-edge-tts

Microsoft Edge Online Text-to-Speech Service (Ruby Gem).

This is a Ruby implementation of the Microsoft Edge Online Text-to-Speech (TTS) service. It provides a simple and easy-to-use interface, allowing you to use high-quality Edge TTS voices directly within your Ruby applications or via the command line.

This project is a port of the Python project edge-tts.

• High-Quality Voices: Direct access to Microsoft Edge's online neural voices.
• Multi-Language Support: Supports 100+ languages and 400+ voices.
• Highly Customizable: Adjustable rate, volume, and pitch.
• Subtitle Generation: Supports generating subtitles in SRT format.
• Command-Line Tools: Includes rb-edge-tts and rb-edge-playback CLI tools.
• Lightweight: Removed unnecessary async dependencies, using standard Ruby libraries and efficient WebSocket handling.

Add this line to your application's Gemfile:

gem 'rb-edge-tts'

And then execute:

$ bundle install

Or install it yourself as:

$ gem install rb-edge-tts

rb-edge-tts provides a powerful command-line interface.

Basic Usage: Generate MP3

$ rb-edge-tts --text 'Hello, world!' --write-media hello.mp3

Generate Audio and Subtitles

$ rb-edge-tts --text 'Hello, world!' --write-media hello.mp3 --write-subtitles hello.srt

Use a Specific Voice

$ rb-edge-tts --text 'Hello, world!' --voice en-GB-SoniaNeural --write-media hello.mp3
$ rb-edge-tts -f test.txt --write-media test.mp3 -v de-DE-AmalaNeural

Adjust Parameters (Rate, Volume, Pitch)

$ rb-edge-tts --rate=+20% --volume=+10% --pitch=+5Hz --text 'Hello, world!' --write-media output.mp3

Instant Playback

Use the rb-edge-playback command to play the generated speech immediately (requires mpv installed):

$ rb-edge-playback --text 'Hello, world!'

List All Available Voices

$ rb-edge-tts --list-voices

Basic Example

require 'rb_edge_tts'

# Use default voice
communicate = RbEdgeTTS::Communicate.new('Hello, world!')
communicate.save("output.mp3")

Advanced Example: Custom Parameters

require 'rb_edge_tts'

communicate = RbEdgeTTS::Communicate.new(
  'Hello, world!',
  "en-US-AriaNeural",
  rate: "+10%",      # Speed
  volume: "+20%",    # Volume
  pitch: "+5Hz"      # Pitch
)
communicate.save("output.mp3")

Generating Subtitles

require 'rb_edge_tts'

communicate = RbEdgeTTS::Communicate.new('Hello, world!')
submaker = RbEdgeTTS::SubMaker.new

File.open("output.mp3", "wb") do |file|
  communicate.stream do |chunk|
    if chunk.type == "audio"
      file.write(chunk.data)
    elsif %w[WordBoundary SentenceBoundary].include?(chunk.type)
      submaker.feed(chunk)
    end
  end
end

File.write("output.srt", submaker.to_srt)

Streaming

require 'rb_edge_tts'

communicate = RbEdgeTTS::Communicate.new('Hello, world!')

communicate.stream do |chunk|
  if chunk.type == "audio"
    # Process audio data chunk (chunk.data)
    print "." 
  end
end

You can use VoicesManager to find and filter available voices.

require 'rb_edge_tts'

# Get all available voices
voices = RbEdgeTTS::VoicesManager.create

# Find all Chinese (Simplified) Female voices
chinese_female_voices = voices.find(locale: "zh-CN", gender: "Female")

chinese_female_voices.each do |voice|
  puts "#{voice.short_name}: #{voice.friendly_name}"
end

This section is for developers who want to contribute to rb-edge-tts or build from source.

• Ruby 3.0 or higher

1. Clone the repository

   git clone https://github.com/ZPVIP/rb-edge-tts.git
   cd rb-edge-tts

2. Install dependencies

   bundle install

3. Run tests

   We use rspec for testing.

   bundle exec rspec

If you modified the code and want to test it locally:

1. Build the Gem

   gem build rb-edge-tts.gemspec

   This will generate an rb-edge-tts-<version>.gem file.

2. Install the Gem

   gem install ./rb-edge-tts-<version>.gem

3. Verify Installation

   rb-edge-tts --version

To publish a new version to RubyGems (requires permissions):

1. Update the version number in lib/rb_edge_tts/version.rb.

2. Update CHANGELOG.md.

3. Build and push:

   gem build rb-edge-tts.gemspec
   gem push rb-edge-tts-<version>.gem

• eventmachine: For WebSocket event loop.
• faye-websocket: WebSocket client implementation.
• json: JSON data processing.
• terminal-table: For formatting CLI output.
• net/http: For fetching voice lists (Standard Library).

This project is licensed under the GNU Lesser General Public License v3.0 (LGPLv3).

The lib/rb_edge_tts/srt_composer.rb file is licensed under the MIT License.

This project is a Ruby port of the Python project edge-tts. Thanks to rany2 for developing the original version, making high-quality TTS conversion possible via the Edge interface.

For questions regarding Python implementation details, please refer to the original project.

Fork on GitHub | Report an Issue