docs: add Speech Swift TTS integration

[ivan-digital]

Summary

Adds a community tutorial for using Speech Swift as a local OpenAI-compatible text-to-speech backend on Apple Silicon. The guide covers:

• Homebrew installation and server startup
• native and Docker Desktop networking
• Open WebUI audio settings
• selectable local model aliases
• direct endpoint testing and troubleshooting

Release prerequisite

The endpoint implementation merged in soniqo/speech-swift#415. The tutorial explicitly states that v0.0.22 does not contain the route and gates the HTTP setup on the first compatible release, an updated Homebrew formula, or a current-source build.

Validation

• All 402 source MDX files compiled successfully.
• The complete Docusaurus production build succeeded.
• Unsupported vibevoice-1.5b was removed from the selectable HTTP alias table; all remaining aliases map to implemented dispatcher cases.
• git diff --check passed.

[ivan-digital]

Speech Swift v0.0.23 is published and the current Homebrew core formula installs v0.0.23. I also verified that the current Open WebUI OpenAI TTS handler reads the returned Content-Type and transcodes WAV responses to MP3, so the documented Speech Swift endpoint is usable as written.