Whisper ASR Box provides a general-purpose speech recognition webservice by wrapping OpenAI Whisper, faster-whisper, and whisperX models. It exposes a REST API for audio transcription and supports both CPU and GPU deployments via Docker containers. The service handles multilingual speech recognition, language identification, and can be configured via environment variables for model selection and caching.