Most piano learning apps grade you on one thing: did you hit the right note at the right time. ROLI's new AI Music Coach is trying something different — it watches how you play, not just whether you played it correctly, and talks you through what to change.

Hand Tracking Meets Conversational AI

AI Music Coach lives inside the ROLI Learn app and is unlocked by Airwave, ROLI's camera-based hand-tracking accessory. Airwave uses infrared cameras to track 27 joints across both hands at 90 frames per second, positioned above a keybed. Combined with a conversational AI layer, the system can see posture, finger placement, and movement in addition to the notes a student plays, then respond with spoken, personalized feedback in real time. According to ROLI's product overview, the coach supports 40 languages and is designed to feel like a conversation rather than a graded exercise.

The feature builds directly on hardware ROLI has been developing for a couple of years now: Airwave itself launched as a standalone gesture-tracking device, reviewed by MusicTech as "hyper-expressive MIDI control in six dimensions", and later paired with ROLI's Seaboard 2 keybed to add a touchless gestural layer on top of MPE's per-note expression. AI Music Coach is the first release to lean specifically into using that hand-tracking data for teaching rather than performance.

Early Impressions

Independent hands-on coverage has been cautiously positive. MusicTech's first look described the tool as feeling "less like a concept demo and more like a product designed by people who actually practice," and MusicRadar's exclusive early test put the obvious question directly to reviewers: can it really substitute for a human teacher? The consensus so far leans toward AI Music Coach being a strong practice companion rather than a full replacement — useful for the hours a student spends alone at the keyboard between lessons, when a teacher isn't in the room to catch a tense wrist or a rushed passage.

Forbes' coverage and Magnetic Magazine both noted that the real differentiator isn't the AI conversation layer on its own — plenty of apps now offer chatbot-style tutoring — it's pairing that layer with actual movement data from Airwave's cameras, so feedback can reference specific physical technique rather than just note timing pulled from MIDI input.

Why It's Relevant Beyond Beginners

AI Music Coach is aimed squarely at learners, but it's worth watching for anyone interested in expressive controllers generally. Airwave's underlying pitch — tracking continuous hand and finger movement as a performance input, not just a teaching tool — is the same idea that makes MPE controllers interesting to advanced players: expression comes from continuous, per-gesture data rather than discrete on/off triggers. A system built to analyze that kind of data for teaching is, in principle, the same kind of system that could eventually feed it into performance mapping.

Getting Started

Per Music Ally's reporting, Airwave hardware runs £299, with ROLI Learn subscriptions starting at £13 a month or £70 a year. As of June 2026, AI Music Coach is available to all existing Airwave owners rather than being limited to a beta group, and ROLI's blog post on the wider rollout frames it as an ongoing feature rather than a one-time launch, with more instrument- and repertoire-specific coaching modules expected to follow.

For a market that's seen no shortage of "AI-powered" learning claims, ROLI's approach stands out mainly for grounding the AI feedback in real physical measurement rather than audio analysis alone — a detail worth watching as camera-based tracking finds its way into more expressive instrument categories beyond the piano.

It also raises an obvious follow-up question for the wider MPE community: if a camera system can reliably parse hand posture and finger independence well enough to coach a beginner, the same data could eventually inform how expressive gestures get mapped to sound for advanced players too, blurring the line between a teaching tool and a genuinely new performance controller.