Emacsvox sends speech through a speech server. For a new installation, use Omnivox. The server communicates with the engine that produces speech: normally eSpeak NG on GNU/Linux, AVSpeechSynthesizer on macOS, and WinRT SpeechSynthesizer on Windows.
Follow Installation if you have not yet obtained speech. This chapter is for choosing a supported platform route, installing or configuring a different backend, and diagnosing problems. Omnivox’s native build procedures, engine behaviour, and versioned protocols belong to the Omnivox repository.
The menu routes here describe ui-experiments. Release packages can provide
an earlier interface; use the matching manual and keep a working installation
available while testing the development branch.