This does the same thing with Apple models or Ollama models. Also has meeting transcription: https://voice.baselinemakes.com