SayaaS - Say as a Service
=========================

Overview
--------

Do you enjoy making computers or phones talk?  Have you ever needed a
text-to-speech engine that doesn't sound terrible?  Do you have an old
Mac lying around somewhere, with nothing to do?

Introducing SayaaS, the Say command as a Service!

The sayaas script is a simple web service wrapper around OS X's say(1)
command.  It is perfect for use as the text-to-speech engine for an
Asterisk phone system, as it is capable of directly creating output that
meets Asterisk's rather strict requirements for WAV file formatting,
while yielding speech that is far more intelligible over the phone than
output from the text-to-speech programs available natively on FreePBX.

Installation
------------

SayaaS is designed to be extremely simple and easy to set up, so it is
a standalone python script with no dependencies outside of the say(1)
command and the python standard library.  Simply copy it wherever it is
needed and run it from there.

For the sake of convenience, this source tree also includes a
launchd.plist(5) file as well as a Makefile that can install sayaas for
use as a system service.

API
---

GET  /
    Display general information about the service.

POST /v1/wav
    Convert the supplied text to speech in WAV format.  Select
    expected by Asterisk.

    Parameters:
        format: format of the audio data to generate, in the
                following format:

                    [BE|LE](F|I|UI)(8|16|24|32|64)@[0-9]+

                This consists of endianness, data type, sample
                size, and sample rate.

                "asterisk" is a shortcut for LEI16@8000, the format
                expected by Asterisk.

        voice:  name of the voice to use

GET  /v1/voices
    Show a list of available voices.