Flagship Scenarios

YH-UI becomes differentiated when it helps teams ship complete product surfaces, not when it simply lists more components.

The first phase should build recognition around three flagship scenarios. Each one should feel like a real product surface, not a disconnected demo fragment.

1. AI Workspace

Target users

• AI SaaS products
• copilot-style tools
• internal assistants
• multi-panel analysis workspaces

Capabilities to showcase

• AiChat, AiBubble, AiConversations
• AiSender, AiEditorSender
• AiThoughtChain
• AiCodeBlock, AiCodeEditor, AiCodeRunner
• AiArtifacts, AiAttachments, AiSources
• streaming, queueing, retry, and stateful flows from @yh-ui/request

Minimum product surface

• left conversation list
• central chat surface
• right artifacts / sources / task panel
• top model and workspace switchers
• realistic error, cancel, retry, and source-expansion states

Current repository status

This scenario has already started landing as the Workspace route inside apps/ai-ops-starter, rather than living only as a planning note.

The route already includes:

• a left conversation rail
• a central generative surface
• a right reasoning and source column
• prompt shortcuts
• an artifact drawer

That gives YH-UI a real product surface for validating AI-component fixes instead of relying only on isolated docs demos.

Acceptance focus

• long conversations remain smooth
• keyboard usage remains complete
• code, attachments, sources, and text coexist cleanly
• streaming and failure recovery feel explicit and trustworthy

2. Flow Automation Console

Target users

• automation platforms
• AI agent orchestration tools
• workflow designers
• internal process builders

Capabilities to showcase

• @yh-ui/flow
• nodes, edges, controls, minimap, and background
• auto layout, import/export, screenshots
• AI node configuration panels
• themed sidebars and multi-state tool surfaces

Minimum product surface

• central canvas
• left node catalog
• right properties / configuration panel
• top run, import, export, and version controls
• viewport controls, zoom, lasso, undo, redo

Acceptance focus

• large canvases remain fluid
• import/export restores useful state
• selection, connecting, zooming, and dragging remain stable
• node configuration and canvas behavior feel like one product

3. Heavy Data Operations Console

Target users

• ecommerce operations teams
• internal admin consoles
• review and audit systems
• multi-form, permission-heavy products

Capabilities to showcase

• YhTable
• virtual scrolling, column drag, resize, fixed columns
• filtering, sorting, pagination, bulk actions
• import, export, print
• FilterBar, Descriptions, and business cards

Minimum product surface

• top filter and toolbar band
• main table with summary information
• drawer / dialog details
• column settings with persisted preferences
• export and print entry points

Acceptance focus

• stable scrolling at large row counts
• persistent column and filter choices
• trustworthy print and export output
• a clear narrow-viewport strategy, not just desktop screenshots

Shared standards

Dimension	Standard
Product feel	Show complete workflows before isolated widgets
Verifiability	Every scenario should be eligible for smoke, a11y, and visual regression checks
Reuse	Layouts and slices from these scenarios should feed starters or blocks
Documentation	Each scenario should document target users, required modules, and acceptance criteria

Scope for this phase

The current phase focuses on:

• scenario definition
• scenario landing pages
• information architecture
• component-composition plans
• acceptance standards

It does not require turning all three scenarios into full standalone products immediately, but later starters and showcase entries should orbit these same product surfaces.