Add python-multipart dependency

Simplify: Remove proxy service, keep only audio file upload for HF Spaces

Add YouTube proxy and audio transcribe endpoint

Add Whisper transcription support and config

Add embeddings API, caching, and Docker revamp

init yt