Shinro Studio

Shinro Studio is the IDE — the operator-facing tool on top of the Cake kernel: five modes, one tool. The unifying theme is ease of use: lowering the barrier to composition, scaffolding modules from vendor docs with an AI assistant, and updating deployed modules without halting the rest of the system.

The five modes

Blueprint

Visual hardware composition. Drop a Jetson on the canvas, attach a LiDAR, wire a SLAM module into it. The shape of the robot becomes a graph you can rearrange.

Flow

Visual behavior composition. Patrol patterns, emergency stops, conditional actions. The behavior of the robot becomes the same kind of graph as its hardware — composable, inspectable, versionable.

Code

When visual composition stops being the right abstraction, an embedded code editor picks up the rest. Customize a module. Author one from scratch. Step through it with the integrated debugger, which is itself a debugger module — swap in GDB, LLDB, or a vendor debugger as the target requires.

Simulator

Today the Simulator is a lightweight 3D viewer — a happy-path visual surface, not a production-grade physics simulation. MuJoCo, Gazebo, Isaac Sim, and others are planned as simulator modules managed by the Cake kernel. Integration with formal verification methods (Mununu and alternatives) and agentic AI workflows (local models or MCP/API integration) is part of the same module-driven simulator roadmap.

A production physics path already exists outside the Studio today: the Python modules architecture runs whole-body control directly against MuJoCo as a standalone application. Wiring that framework into Studio’s Simulator mode as a managed module is tracked on our roadmap; it’s not yet built.

Deployment Manager

Compile, validate, push to target hardware, hot-swap individual modules. The Cake kernel is resident on the robot to load and swap modules, but it does not interpose on the data those modules exchange. This is the hardest component of the Studio, and currently partial.

Why this matters

The Blueprint Editor lowers the barrier to composition. The AI assistant scaffolds modules from vendor docs. Hot-swap means a deployed module can be updated without halting the rest of the system. Together, they move composition, scaffolding, and deployment into a single operator workflow.