feat: enable OpenAI Realtime STT for streaming speech recognition

Switch from batch STT (gpt-4o-transcribe via /audio/transcriptions)
to streaming Realtime API (WebSocket). This eliminates the ~2s batch
upload+process latency per utterance.

Also updated nginx proxy on 67.223.119.33 to support WebSocket upgrade
for /v1/realtime endpoint.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

packages/services/voice-agent/src/agent.py

@@ -148,6 +148,7 @@ async def entrypoint(ctx: JobContext) -> None:
             model=settings.openai_stt_model,
             language=settings.whisper_language,
             client=_oai_client,
+            use_realtime=True,
         )
     else:
         stt = LocalWhisperSTT(