Krea adds live voice control to iPad canvas
Krea's new Voice Mode lets iPad users speak edits while they draw, and the canvas updates in real time with no prompt bar, submit button or break from sketching.
Creative-AI company Krea has launched Voice Mode, a feature that lets users speak instructions to its image model while drawing on an iPad. The feature is available now in Krea for iPad.
To use it, users open the canvas and tap a microphone icon, then talk while they draw. Krea's examples include "make the background warmer" and "add more contrast to the left side," said while the pencil is still on the screen. According to the company, the model reads each request in the context of the drawing and applies the change live. There is no separate prompt bar or submit step.
Krea says the response is continuous, so users can adjust as they go. Saying "darker" shifts the image, and a follow-up "not that much" pulls the change back. The company also describes combining touch with speech, such as pointing at a region and saying what should change, or sketching a rough shape and describing its material.
Krea says it launched on iPad first because a tablet is the closest thing to working on paper. On a desktop, the user's hands are already on a keyboard, which makes voice less of a natural fit.
The announcement leaves several things out. It does not name the underlying model, give latency figures or explain how voice input is processed. It also gives no timeline for other platforms or for availability beyond the iPad app.