As a user of a Qt application, I want to be able to trigger specific application actions by voice command.

This would typically be a sequence that starts with a wake-up word (like "Hey Siri" or "Alexa"), and then a natural language representation of a well-defined activity that is specific to the context of the current application or domain ("make the current paragraph bold").

This is different from free text dictation via speech recognition, which is usually covered by the platform's input method framework.

Open Questions

Should an application inform the engine about what actions there are and what valid phrases it can support?

Such a system would need to be backed by some natural language processing machinery, ie. mapping "Hey TextEdit, make the current paragraph bold" to a sequence of actions (Select current paragraph, select the "bold" action, then remove the selection). How much of that is already provided by each platform's speech recognition framework?

• macOS
• iOS
• Windows
• Android
• Linux