Audio-transcription capability

Trussium exposes provider-neutral, non-streaming audio transcription at POST /v1/audio/transcriptions.

The endpoint receives multipart form data and forwards the audio directly to the configured provider. Trussium does not persist the audio, transcript, credential, endpoint, or provider payload. Structured logs receive only the normal request, execution, capability, provider, and model context.

curl http://localhost:9000/v1/audio/transcriptions \
  -F model=gpt-4o-transcribe \
  -F language=en \
  -F file=@recording.wav

{
  "id": "bfc37f15-ff22-4b57-a8f2-aa01ab9af121",
  "provider": "openai",
  "model": "gpt-4o-transcribe",
  "text": "Hello world",
  "language": "en",
  "duration": 1.2,
  "segments": [
    {"id": 0, "start": 0.0, "end": 1.2, "text": "Hello world"}
  ]
}

The initial capability supports direct single-request transcription and normalized segment metadata when supplied by the provider. Streaming transcription, translation, diarization, audio storage, asynchronous jobs, routing, retries, and fallback remain separate concerns.