AI Workflow Builder vs Headroom

This demo shows how JointJS for React handles agent-building UIs: adding skills and prompts, wiring up tools, and defining agent behavior visually.

It also checks your driver for the subgroup bug that makes some in-browser LLMs produce garbage.

About AI Workflow Builder

This demo shows how JointJS for React handles agent-building UIs: adding skills and prompts, wiring up tools, and defining agent behavior visually. It ships with automatic layout, custom shapes, a navigator and many more features out of the box, the same pieces you'd use to build this into your own product. Extend the default workflow or build your own from scratch, add new actions, configure agents, jot down notes, and explore everything the diagram can do.

About Headroom

It also checks your driver for the subgroup bug that makes some in-browser LLMs produce garbage. One outbound request: ~0.4 MB of public model metadata to derive bytes per token, no run data attached. Projections are measured read bandwidth ÷ bytes touched per token, at empty context; derived manifests make bytes-per-token exact, estimates fall back to an honest band. Real engines land below the ceiling; the gap is engine overhead (see methodology). Runs the transformers.js WebGPU path on the exact model your ceiling is quoted for ( onnx-community/gemma-4-E2B-it-ONNX.