Skip to content
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
194 changes: 175 additions & 19 deletions ai/agentic-ui-generator/overview.md
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
title: Overview
page_title: Telerik Blazor MCP Server Overview
description: Learn about the Telerik Blazor MCP Server, its modes of operation, and how it enhances AI-powered development with Telerik UI for Blazor components.
title: Telerik UI for Blazor AI Tools Overview
page_title: Telerik UI for Blazor AI Tools Overview
description: Learn how Telerik UI for Blazor AI Tools help you generate pages, configure components, and choose the right MCP-assisted workflow.
slug: ai-overview
tags: ai, mcp, assistant, agentic, generator
published: True
Expand All @@ -14,7 +14,40 @@ tag: updated

The Telerik UI for Blazor AI Tools are delivered through a single [Model Context Protocol (MCP) server](https://modelcontextprotocol.io/docs/getting-started/intro) that connects your AI client to UI-generation capabilities and knowledge specific to Telerik UI for Blazor.

From idea to implementation, you can use the MCP server to generate complete pages, configure components correctly, align with the Progress Design System, and reduce repetitive setup work.
Use the MCP server to generate complete pages, configure components correctly, align with the Progress Design System, and reduce repetitive setup work. This overview helps you choose between the full Agentic UI Generator flow and the specialized assistants for setup, layout, component configuration, styling, icons, accessibility review, and validation.

## Before You Start

This overview assumes that you have:

* An AI client that supports MCP servers.
* Either an existing Telerik UI for Blazor project or a new project that you want to scaffold during setup.
* Access to an active Telerik subscription or a trial when you want to use the AI tools end to end.

Use this page to decide which AI tool fits your task. For installation steps, client-specific setup, and troubleshooting, continue with [Agentic UI Generator Getting Started](slug:agentic-ui-generator-getting-started).

If your client can connect to MCP servers but uses a different prompt UX, the same assistant roles still apply even when the exact interaction flow looks different.

## Supported AI Clients and .NET IDEs

The Telerik Blazor MCP Server is designed for AI clients that support MCP server integrations. The client provides the chat experience and sends requests to the Telerik tools; the Agentic UI Generator and specialized assistants provide Telerik UI for Blazor context.

The documentation includes setup guidance for the following AI clients:

| AI client | Compatibility status | Notes |
|---|---|---|
| GitHub Copilot | Documented for MCP-compatible hosts | Use the MCP configuration and enable the Telerik tools in the host's tool selection when required. |
| Cursor | Documented | Add the Telerik MCP server to the workspace or user configuration as described in the setup article. |
| Claude Code | Expected when MCP server support is available | Follow the client's current MCP configuration guidance. The client vendor controls its configuration and feature availability. |
| Windsurf | Expected when MCP server support is available | Follow the client's current MCP configuration guidance. The client vendor controls its configuration and feature availability. |

You can use the Telerik Blazor MCP Server from common .NET development environments, including:

* **Visual Studio**
* **Visual Studio Code**
* **JetBrains Rider**

The exact configuration steps depend on the IDE and AI client combination. For current setup examples, see [Agentic UI Generator Getting Started](slug:agentic-ui-generator-getting-started). Compatibility outside the documented examples depends on the client's MCP implementation and should be verified in the target environment.

## What Are the Telerik UI for Blazor AI Tools

Expand Down Expand Up @@ -51,6 +84,7 @@ The Telerik Blazor MCP server uses an orchestration-first model, centered on the
<Column count={[24,12,8]}>
<Component className="tile card-icon" href="#styling-assistant">
<ComponentTitle>Styling Assistant</ComponentTitle>
</Component>
</Column>
<Column count={[24,12,8]}>
<Component className="tile card-icon" href="#accessibility-assistant">
Expand All @@ -70,62 +104,184 @@ The Agentic UI Generator orchestrates all assistants so you can build pages and

## How the Agentic Flow Works

The Agentic UI Generator takes one prompt and manages the flow for you. It decides which assistants to use and combines their output into a single result. Use it when you want to generate a full page quickly, or call a specific assistant when you need a focused update to the layout, components, styling, theme, or icons in your project.
The Agentic UI Generator takes one prompt and manages the flow for you. It decides which assistants to use and combines their output into a single result. Choose this path when you want to generate a full page quickly, keep several design and implementation decisions coordinated, or move from prompt to first draft without manually switching tools.

![Full Pages](../images/ui-templates.png)

## Which Path to Choose

Use the following scenarios to decide whether you need orchestration or a single assistant:

| Scenario | Best choice | Why |
|---|---|---|
| Generate a dashboard or landing page from one prompt | `#telerik_ui_generator` | The orchestrator can coordinate layout, components, styling, icons, accessibility, and validation in one pass. |
| Set up a new project and connect the MCP server | Getting Started Assistant | It focuses on project scaffolding, MCP configuration, Telerik UI for Blazor setup, and license activation. |
| Refine page hierarchy, spacing, and responsive breakpoints | Layout Assistant | It is optimized for section structure and responsive presentation decisions. |
| Configure a Grid, form, or other Telerik UI for Blazor component | Component Assistant | It helps with component selection, API usage, and feature configuration such as filtering, grouping, or export. |
| Apply brand tokens, theme variables, or high-contrast adjustments | Styling Assistant | It focuses on reusable styling decisions instead of component behavior. |
| Review keyboard flow, ARIA usage, and semantic markup before release | Accessibility Assistant | It is the right tool for WCAG-oriented review and corrective guidance. |

## What You Get Back

The result depends on the assistant that you call and the prompt that you provide. Typical outputs include:

* Proposed page structure and section hierarchy for a new screen.
* Telerik UI for Blazor component recommendations and configuration guidance.
* Generated or revised markup, styling guidance, and theme-token suggestions.
* Accessibility corrections for keyboard flow, ARIA usage, semantic markup, and contrast issues.
* Validation feedback that helps you review whether the generated result follows Telerik UI for Blazor best practices.

Treat the generated result as an accelerated first draft. Review data binding, business rules, service integration, and project-specific architecture before you ship the implementation.

## Prompt Examples

The following examples show when to choose the orchestrator or a targeted assistant:

| Goal | Best choice | Example prompt | Expected result |
|---|---|---|---|
| Create a complete analytics dashboard page | `#telerik_ui_generator` | `Create a Blazor analytics dashboard with KPI cards, a revenue chart, a filter form, and a pageable Grid for recent orders.` | A coordinated first draft that can include layout, Telerik UI for Blazor component choices, styling direction, and validation feedback. |
| Configure a Grid for richer behavior | Component Assistant | `Configure a Telerik UI for Blazor Grid with filtering, grouping, export, and virtual scrolling for an orders dataset.` | Focused guidance for Grid features, relevant configuration choices, and API-oriented implementation direction. |
| Refine layout without changing component behavior | Layout Assistant | `Rework this page into a two-column tablet layout and a single-column mobile layout while preserving the existing components.` | Revised section hierarchy, spacing, and responsive layout guidance. |

Use targeted assistants when the scope is narrow. Use the orchestrator when one request spans layout, components, styling, accessibility, and validation.

## How a Real Workflow Looks

The following examples show what an end-to-end interaction can look like in practice.

### Orchestrated Workflow Example

**Scenario:** You have a Blazor line-of-business app and need a new sales dashboard page.

**Prompt:** `Create a sales dashboard page with KPI cards, a revenue trend chart, a filter form, and a Telerik UI for Blazor Grid that shows recent orders with paging and filtering.`

**What the orchestrator typically coordinates:**

* Page structure and section order.
* Telerik UI for Blazor component selection for the cards, chart, form, and Grid.
* Styling direction that keeps the page aligned with the design system.
* Validation feedback for generated structure and Telerik usage patterns.

**What you still review manually:**

* Real data binding and service calls.
* Domain-specific filtering rules and business calculations.
* Security, authorization, and application-specific error handling.

### Targeted Workflow Example

**Scenario:** The page already exists, but the Grid configuration is incomplete.

**Prompt:** `Update this Telerik UI for Blazor Grid to support grouping, export, and virtual scrolling for large order datasets.`

**What the Component Assistant is best at:**

* Narrow component-level guidance instead of full-page regeneration.
* Feature-oriented configuration suggestions for the Grid API.
* Safer iteration when you want to refine one part of the page without changing layout or styling.

**What it does not replace:**

* Your application data model and transport layer decisions.
* Performance testing with your real dataset size and hosting constraints.

## Example Workflows

The following workflows show where the tools add Telerik-specific value:

1. First-time setup and first generated page:
1. Start with the Getting Started Assistant to scaffold the project, connect the MCP server, and activate licensing.
2. Prompt `#telerik_ui_generator` to build the first page.
3. Let the orchestrator route layout, component, styling, and validation work automatically.
2. Existing application, new dashboard feature:
1. Ask the Layout Assistant to propose the section structure for the dashboard.
2. Use the Component Assistant to configure the Grid, charts, forms, or navigation elements.
3. Bring in the Styling Assistant to apply tokens and theme refinements.
3. Final review before handoff:
1. Run the Accessibility Assistant to catch keyboard, ARIA, and contrast issues.
2. Use the Validator Assistant through the orchestrator to confirm Telerik UI for Blazor best practices.
3. Return to a targeted assistant only when one part of the page needs refinement.

### Getting Started Assistant

Use the Getting Started Assistant when you want a guided onboarding flow for first-time setup. Call it to scaffold a Blazor project, configure the Telerik MCP server, set up Telerik UI for Blazor in the project, and complete license activation with minimal manual steps.
Starting from a blank machine or a clean proof-of-concept is the best fit for the Getting Started Assistant. It guides the first-run experience: project scaffolding, MCP configuration, Telerik UI for Blazor setup, and license activation.

It is useful when setting up a new environment, validating your initial MCP integration, or preparing a clean proof of concept quickly.
Best for: first-run setup, clean environment validation, and early proof-of-concept work.

It reduces setup friction, but it does not replace feature-specific documentation after the environment is ready.

### Layout Assistant

Use the Layout Assistant to set up or refine the page structure. It helps with section order, spacing, and responsive behavior so the UI stays clear across desktop, tablet, and mobile.
Layout work becomes easier once you know the content and component mix, but need help shaping the page around them. The Layout Assistant focuses on section order, spacing, and responsive behavior so the UI stays clear across desktop, tablet, and mobile.

Best for: dashboard structure, responsive adaptation, and visual hierarchy cleanup.

Typical tasks include adding a new dashboard section, cleaning up visual hierarchy, and converting desktop-first screens into responsive layouts.
Typical outcomes include a clearer dashboard hierarchy, better spacing between sections, and responsive layouts that do not collapse awkwardly on smaller screens.

![Layout Assistant](../images/layout-assistant.png)

### Component Assistant

Use the Component Assistant when you need help configuring Telerik UI for Blazor components. It helps you pick the right component and wire it correctly with real API patterns.
Component work is where the product-specific guidance is most visible. The Component Assistant helps you choose the right Telerik UI for Blazor component and configure it with real API patterns.

Common tasks include enabling Grid features (sorting, paging, filtering, grouping), building validated forms, setting up virtual scrolling or export, and using sample data for safe prototyping.
Best for: component selection, Telerik API usage, and feature configuration for controls such as the Grid and form inputs.

It is a strong fit for Grid features such as sorting, paging, filtering, grouping, export, and virtualization, as well as validated forms and safe sample-data prototypes.

![Component Assistant](../images/component-assistant.png)

### Styling Assistant

Use the Styling Assistant when you want consistent visuals across the app. It helps define reusable tokens and CSS variables for scalable theming.
Styling decisions pay off early when you want more than a one-off visual tweak. The Styling Assistant helps define reusable tokens and CSS variables for scalable theming.

Best for: theme tokens, reusable CSS variables, and consistent visual refinement across multiple pages.

Typical tasks include applying brand colors, adding dark mode or high-contrast variants, and keeping styling behavior consistent as new pages are added.
Bring it in for brand color systems, dark mode or high-contrast variants, and reusable styling rules that can scale as new pages are added.

![Styling Assistant](../images/style-assistant.png)

### Icon Assistant

Use the Icon Assistant to choose icons that match user actions and UI context. This assistant helps you achieve visually consistent navigation, status indicators, and action buttons.
Icons are a small surface, but inconsistency shows up quickly in navigation and actions. The Icon Assistant helps you choose icons that match user intent and UI context.

Best for: action icons, navigation cues, and status indicators that need visual consistency.

It is useful for toolbars, navigation menus, cards, and any new section where icon consistency matters.
It is especially useful for toolbars, navigation menus, cards, and status indicators where visual meaning must stay consistent across the app.

![Icon Assistant](../images/icon-assistant.png)

### Accessibility Assistant

Use the Accessibility Assistant to apply WCAG 2.2 Level AA guidance during implementation, not after it. It helps with ARIA usage, keyboard navigation, semantic markup, and color contrast validation for text and UI controls.
It is especially useful for interactive templates, complex component flows, and final semantic checks before release.
Accessibility review is most effective before the page structure is locked down. The Accessibility Assistant applies WCAG 2.2 Level AA guidance during implementation and helps with ARIA usage, keyboard navigation, semantic markup, and color contrast validation for text and UI controls.

Best for: interactive templates, keyboard review, semantic corrections, and accessibility checks before release.

It is most valuable for interactive templates, complex component flows, and final semantic checks before release.

![Accessibility Assistant](../images/accessibility-assistant.png)

### Validator Assistant

Not designed to be invoked manually. It is called automatically by the UI Generator Orchestrator and ensures the generated code follows Telerik UI for Blazor best practices and standards.
The Validator Assistant is not a manual entry point. The orchestrator calls it automatically to check whether generated output follows Telerik UI for Blazor best practices and standards.

It helps confirm component usage, generated structure, and implementation direction, but it does not replace manual review of business logic, data access, or application-specific requirements.

### When to Use Orchestrated vs Targeted Mode

Use `#telerik_ui_generator` for a complete orchestration-first workflow from a single prompt. When you need finer control or want to adjust just one aspect (such as layout, theme, or a component), you can call a specialized assistant directly by its dedicated handle. For details, see [Target the Assistants (Advanced)](slug:agentic-ui-generator-prompt-library#assistant-specific-prompts).
Use `#telerik_ui_generator` when one request spans layout, components, styling, accessibility, and validation. Switch to a dedicated assistant when the page already exists and you want to refine a single area without rerunning the full orchestration flow. For details, see [Target the Assistants (Advanced)](slug:agentic-ui-generator-prompt-library#assistant-specific-prompts).

## Limitations and Unsupported Scenarios

Keep the following boundaries in mind:

* This overview does not replace the detailed setup, troubleshooting, or API-focused articles linked from Getting Started and the Prompt Library.
* The AI tools can generate and refine UI, but application-specific business logic, backend integrations, and data contracts still require manual review and implementation decisions.
* The Validator Assistant is called through the orchestrator rather than used as a separate manual workflow.
* The output quality depends on the prompt quality, the available project context, and the capabilities of the client that connects to the MCP server.
* Existing-project results are only as complete as the context that the client can send. Missing component markup, incomplete data-model context, or partial file visibility can reduce output accuracy.
* Generated output can accelerate page composition and component setup, but it does not automatically verify production readiness, application architecture, or environment-specific deployment constraints.
* Some prompts need iterative refinement, especially when the request depends on custom data models, nonstandard design constraints, or project-specific architecture.
* License state, local tooling, and AI client capabilities can change the exact setup path for the same overall workflow.

## AI Plugin

Expand Down