Local Speech-to-Text (STT) using Whisper ASR inference snap
This snap provides a local speech-to-text (STT) solution using OpenAI's Whisper models.
It uses the WhisperLive for real-time transcription and provides an OpenAI adapter for exposing an OpenAI-like Realtime Transcription API.
This currently supports CPU inference only. Support for GPU inference is planned for the future.