Real-time speech-to-text transcription using OpenAI’s Whisper Large v3 model via WebSocket with compressed audio support. Supports streaming audio input with intelligent Voice Activity Detection (VAD), partial transcripts for immediate feedback, and automatic language detection.