Speech Dispatcher Service

Summary: Text-to-speech

A tiny service that can speak under WebSocket control. This communicates using a speech-dispatcher instance using the SSIP protocol.

Protocol

topic payload
speech/command/speak { utterance: 'Hello, world!' }
Optionally, voiceType: 'ID' can change the voice
speech/event/spoken { utterance: 'Hello, world!' }
speech/command/list-voices { voices: ['MALE1', 'FEMALE1'] }

Installation

Requires node.js 8+

On Debian-based systems:

sudo apt-get install speech-dispatcher

Then:

npm install --production

Running

WS_HOST=raspberrypi.local WS_PORT=8000 node src/main

WS_PORT is mandatory. The host IP is used if WS_HOST is not supplied.

Debugging

To log out debugging information, set the DEBUG environment variable i.e.

WS_PORT=8000 DEBUG=* node src/main

debug is used for debugging.