TTS provider: OpenAI gpt-4o-mini-tts

[thefourcraft]

Part of the Read Aloud (TTS) epic. Lowest-friction option.

• Reuse existing openai key (Bearer).
• POST /v1/audio/speech, response_format=pcm (24kHz/16-bit mono), chunked streaming.
• Voices: alloy, ash, ballad, coral, echo, fable, nova, onyx, sage, shimmer, verse, marin, cedar; expose free-text instructions field for tone.

File: TextToSpeech/Providers/OpenAITTSProvider.swift.