Abstract: The piano-roll has been the de-facto standard representation for melodic and harmonic content in DAWs for decades, yet direct manipulation of those requires expert knowledge of music theory to begin with, and additionally becomes physically impractical when switching to the smaller, touch-based screens of modern mobile devices, too coarse for the precision required by micro-timings and the unforgiving discrete placement of pitches. Imbuing these interfaces with machine learning and offloading their precise and error-prone aspects to style-adaptive AI assistants may allow the design of more intuitive interactions whilst maintaining a high level of control, helping lower the cost of entry to composition for novices and offer stimulating new creative tools for professional musicians. We introduce PIANOTO, a touch-ready, responsive web interface for creating expressive piano performances through AI-assisted inpainting, all via simple swipe operations. This open-source, model-agnostic prototype is designed with both novice and expert users in mind, for usage either as a standalone tool or in conjunction with existing DAWs, on desktop or mobile.