Back to the Lab
2026 · App · WIP

Bone Maestro

Exploring how to get more control over an animated skeleton from Home Depot.

  • iOS
  • SwiftUI
  • Hardware + Bluetooth
  • AI-assisted build
  • Experimental UX
Hero image for Bone Maestro

A side project exploring how far a consumer Halloween animatronic can go when you redesign the control layer — mobile UX, Bluetooth, custom scenes, and careful build practice.

Why I wanted to build this?

Blueprint-style illustration of a skeleton and a house
Motivation sketch — replace with final art

Me and my family are obsessed with Halloween, and I love product experimentation. The Home Depot Ultra Skelly is fun out of the box — but the stock app stops short of what the hardware feels capable of.

I wanted to push that animatronic past default modes into something more expressive and programmable: clearer controls, custom performances, and room to experiment with voice and character.

What I explored

Four phone screens exploring Scenes, Audio, Voice AI, and Lighting & Motion controls
Early control surfaces for scenes, audio, voice, and motion
  • Designing clearer mobile controls for motion, audio, lighting, and related states.
  • Learning the skeleton’s Bluetooth behavior with a sniffer and a community web controller — treating every command as a candidate until verified.
  • Testing custom scenes for programmable performances.
  • Exploring ideas for AI voice, text-to-speech, and conversational modes (roadmap, not shipped).
  • Documenting technical research and protocol notes along the way.

What I built

Architecture diagram from iOS SwiftUI app through Bluetooth to Ultra Skelly audio, lighting, and motion
App → Bluetooth → skeleton control stack

Bone Maestro is a SwiftUI-based iOS app for controlling the Home Depot Ultra Skelly. The focus is a clearer control surface for Bluetooth connection, animatronic status, and the core performance channels: audio, lighting, and motion.

The app supports custom scenes, volume control, mouth sync, and light controls — treating the skeleton less like a one-shot gag and more like a programmable performer.

Bluetooth connect flow
Live control demo

What I learned

  1. 01Design for the real world.
  2. 02Test, learn, document.
  3. 03AI as a multiplier.
  • Hardware UX needs more clarity than software-only products — connection state, latency, and physical feedback have to stay visible.
  • Experimentation works better when knowns are separated from assumptions, then tested and documented.
  • AI tools help most when paired with strong design judgment, not when they replace it.

Links