Interaction with digital content has changed substantially with the introduction of Augmented Reality (AR). User input in AR is essential for messaging or note-taking and typically relies on mid-air keyboards and controllers. However, these methods lead to fatigue, reduce visual attention, and are not designed for use in public spaces, making them impractical in dynamic contexts, e.g., walking. In this paper, we introduce knuckle-based input in AR for text entry. To explore this idea, we conducted a controlled experiment (N=18) comparing it to the state-of-the-art controller-based input while sitting, standing, and walking. We found that the knuckle method provided a better user experience and was preferred for walking due to its engaging nature and reduced visual focus. In contrast, controller input was favored for sitting and standing due to accuracy and ease of use.