There is an option to set a manual start behavior in the "Mixed Reality Speech Commands Profile"s. How do I access it and toggle the speech input? Documentation gives no information about this.
I can disable and enable the eyetracker like this:
CoreServices.InputSystem.EyeGazeProvider.IsEyeTrackingEnabled
Though I cannot find an option to do so for the SpeechHandler.
You can try to manually get this DataProvider and then enable it. You can refer to the code below and this document - WindowsSpeechInputProvider Class (Microsoft.MixedReality.Toolkit.Windows.Input) | Microsoft Learn.
CoreServices.GetInputSystemDataProvider<WindowsSpeechInputProvider>().StartRecognition();