FrameworksOfficial

shinygenui 0.2.0 brings catalog-constrained generative UI to R Shiny and delegates streaming, cancellation, and attachments to shinychat

CRAN published shinygenui 0.2.0 on September 22. An R Shiny application can collect predefined components with typed arguments into a catalog, then let a model add, update, or remove canvas components through tool calls. The release delegates streaming, cancellation, and attachments to shinychat's shared session controller and retains a trace of successful interface changes for replay without a model.

Why it matters

This gives the R ecosystem, especially data applications and analytical dashboards, a constrained generative-UI path. The model supplies data and component choices while the host retains component implementation and argument validation. Trace and replay also turn generation into an inspectable sequence of UI operations rather than something that can only be recreated by calling the model again.

A closer look

`genui_component()` defines components and arguments, `genui_catalog()` collects the catalog, and `genui_canvas()` with `genui_server()` lets the model add, update, and remove components in a running application. Every model call passes through `genui_dispatch()` first; failures are returned to the model while the Shiny session keeps running. Version 0.2.0 requires shinychat 0.5.0 and ellmer 0.4.1 and moves streaming, cancellation, and attachments to `shinychat::chat_server()`. `genui_trace()` and `genui_replay()` can rebuild the canvas from successful calls without contacting a model. CRAN's November 6 check page reports OK for version 0.2.0 across its listed Linux, macOS, and Windows environments.

Who this is for

Development, data-science, platform, and security teams building constrained generative UI for R, Shiny, data applications, or analytical dashboards.

Original sources

Revisions & corrections (1)

— Previously uncatalogued official CRAN release with a confirmed publication date, versioned change record, implementation documentation, explicit limitation, and current platform check results.

Related events

Frameworks

AG-UI LangGraph 0.0.46 reads native interrupt schemas, unwraps serialized tool results, and declares protocol version at run start

AG-UI released @ag-ui/langgraph 0.0.46 on November 13. The adapter reads Interrupt.response_schema on LangGraph 1.2.12 and later while retaining value-level fallbacks. It also unwraps outputs from serialized LangChain ToolMessage kwargs, skips tool-end events without a call ID, and declares protocolVersion on RUN_STARTED.