Output audio cues in live speech.

Turned out I was overcomplicating things, all I needed to do was use file paths not URIs.
eSpeak NG handled it from there.

Don't delete downloads, was happening before use.

Allow disabling of URL rewriting for SSML output.

Download CSS assets, including new useragent audio cues.

Audio cues are public domain from freesound.org and archive.org via elementary OS's sound theme.

Export C APIs from files defining underlying functions.

Remove dead imports.

Draft C Bindings, so C APIs can be used for I/O.