Batch audio transcription using Whisper inference

Summary

Transcribes up to 500 audio files in a single API call using local Whisper inference. This service accepts an array of audio URLs and optionally specifies the target language.