wsay

Windows equivalent of macOS say.

Simple command line text-to-speech with easy file output, voice selection and more.

Features

Output to wav file.
Supports playback from text files, clipboard and pipes.
Select different voices (including new Windows 10 voices).
Supports speech xml.
Supports selecting playback device.
Supports multiple simultaneous playback devices and file output.
Full unicode (utf8) support (supports asian languages, Slovak, etc).
- Note : Window's legacy command prompt has some of it's own issues with utf8.
- If you need full unicode support, best to use the new Windows Terminal.
Supports utf8, utf16le and utf16be text files.
Interactive mode.
Volume, speed and pitch options.
Radio effects for flight sims and such.

Install Instructions

Copy wsay.exe to a folder you like.
Add that folder to your Path environment variable.
Type wsay in your command prompt.

Examples

# Say something.
wsay "Hello there."

# Ouput to a wav file. If no filename is entered, outputs to 'out.wav'.
wsay "I can output to a wav file." -o

# List supported voices. Install new Windows voices for more choices.
wsay --list_voices
	1 : Microsoft David Desktop - English (United States)
	2 : Microsoft Hazel Desktop - English (Great Britain)
	3 : Microsoft Zira Desktop - English (United States)
	4 : Microsoft David - English (United States)
	5 : Microsoft James - English (Australia)
	6 : Microsoft Linda - English (Canada)
	# etc.

# Use the number provided by '--list_voices' to select a different voice.
wsay "I can use different voices." --voice 6

# Provide an output filename.
wsay "You can name the output file." -o my_output_file.wav

# Read text from a text file instead.
wsay -i i_can_read_a_text_file.txt

# Read text currently copied to clipboard.
wsay -c

# Pipe text.
echo "Speaking from pipe." | wsay

# Pipe in a file.
wsay < "i_can_read_a_text_file.txt"

# When using speech xml on the command line, escape double quotes with backslashes.
wsay "Lets take a little <silence msec=\"500\"/> pause."

# In interactive mode, type sentences and press enter for them to be read.
# Use !exit to quit.
# Use !stop to stop speaking.
wsay --interactive
[Info] Type sentences, press enter to speak them.
[Command] '!exit' : Exit interactive mode.
[Command] '!stop' : Interrupt speaking.

# List supported playback devices.
wsay --list_devices
	1 : BenQ GW2765 (NVIDIA High Definition Audio)
	2 : Speakers (AudioQuest DragonFly Black v1.5)
	3 : Digital Audio (S/PDIF) (High Definition Audio Device)
	# etc.

	Use 'all' to select every device.

# Speak using a non-default playback device.
wsay "I am speaking on another playback device." --playback_device 2

# You can play on multiple devices and save to a file simultaneously.
# Seperate the device numbers with spaces.
# WARNING : The "sentence" must come before the --playback_device (-p) option if it is used!
wsay "Output EVERYWHERE \o/" -p 1 2 -o

# Use the 'all' shortcut to playback on every device simultaneously.
wsay "I conquer all devices" -p all

# You can set the voice volume, from 0 to 100.
wsay "Softly speaking" --volume 25

# You can set the voice speed, from 0 to 100. 50 is the default speed.
wsay "Quickly speaking" --speed 75

# You can set the voice pitch, from 0 to 20. 10 is the default pitch.
wsay "Low voice." --pitch 0
wsay "Normal voice." --pitch 10
wsay "High voice." --pitch 20

# If speaking code or text with special characters, you can disable speech xml parsing.
wsay "<html>Definitely Facebook</html>" --nospeechxml

# Degrade the sound to give a CB / radio feel. Multiple effects are provided.
wsay "3 3 3 Lima Delta, do you know how to operate the transponder?" --fxradio 1
wsay "3 3 3 Lima Delta, if able input 7 7 0 0 into your transponder." --fxradio 6

# Without white-noise.
wsay "3 3 3 Lima Delta, do you know how to operate the transponder?" --fxradio 2 --fxradio_nonoise

# Longer or shorter pauses between paragraphs. Use milliseconds.
(echo "No" & echo."pause.") | wsay --paragraph_pause 0
(echo "Long" & echo."pause.") | wsay --paragraph_pause 1000

# Here, we are using voice 6, reading text from a file and outputting to 'output.wav'.
wsay -v 6 -i mix_and_match_options.txt -o output.wav

# Ouput to multiple devices using interactive mode with voice 5.
wsay -v 5 -I -p 1 2

# Speak slowly and quietly, on all devices, using voice 7 and save to wav file.
wsay "Multiple options example." -v 7 -p all -s 25 -V 25 -o

Help

Usage: wsay "sentence" [options]

Arguments:
 "sentence"    Sentence to say. You can use speech xml.

Options:
 -c, --clipboard                   Speak currently copied text (the text in your clipboard).
 -i, --input_text <value>          Play text from '.txt' file. Supports speech xml.
 -I, --interactive                 Enter interactive mode. Type sentences, they will be spoken when you press enter.
                                   Use 'ctrl+c' or type '!exit' to quit.
 -d, --list_devices                List detected playback devices.
 -l, --list_voices                 Lists available voices.
 -o, --output_file <optional>      Outputs to wav file. Uses 'out.wav' if no filename is provided.
 -P, --pitch <value>               Sets the voice pitch, from 0 to 20. 10 is the default pitch.
 -p, --playback_device <multiple>  Specify a playback device. Use the number provided by --list_devices.
                                   You can provide more than one playback device, seperate the numbers with spaces.
                                   Use 'all' to select all devices.
 -s, --speed <value>               Sets the voice speed, from 0 to 100. 50 is the default speed.
 -v, --voice <value>               Choose a different voice. Use the voice number printed using --list_voices.
 -V, --volume <value>              Sets the voice volume, from 0 to 100.
 -h, --help                        Print this help

Extra Options:
     --fxradio <value>             Degrades audio to make it sound like a radio, from 1 to 6.
     --fxradio_nonoise             Disables background noise when using --fxradio.
     --nospeechxml                 Disable speech xml detection. Use this if the text contains special characters that
                                   aren't speech xml.
     --paragraph_pause <value>     Sets the amount of pause time between paragraphs (in milliseconds), from 0 to *a big
                                   number*.

wsay
version 1.6.2
https://github.com/p-groarke/wsay/releases
Philippe Groarke <hello@philippegroarke.com>

Build

Install recent conan, cmake and compiler.

Windows

mkdir build && cd build
cmake .. && cmake --build .