Telerik Blazor MCP Server
The Telerik Blazor MCP Server lets you interact with AI and reach new levels of developer productivity. The MCP server provides proprietary context to AI-powered IDEs, apps and tools. You can use the Telerik Blazor MCP server for Blazor AI code generation and ask about Telerik UI for Blazor components, features, or general usage. You can successfully prompt more complex questions and tasks, and generate tailored code that includes Telerik UI for Blazor components and API.
Prerequisites
To use the Telerik Blazor MCP server, you need:
- A compatible MCP client (IDE, code editor or app) that supports MCP tools.
- A Telerik user account.
- An active DevCraft or Telerik UI for Blazor license or a Telerik UI for Blazor trial.
- A Blazor application that includes Telerik UI for Blazor.
Installation
Use the documentation of your AI-powered MCP client to add the Telerik MCP server to a specific workspace or globally. The sections below provide installation tips and examples for some popular MCP clients like Visual Studio, VS Code, and Cursor. The generic settings of the Telerik Blazor MCP server are:
- npm package name:
@progress/telerik-blazor-mcp
- Type:
stdio
(standard input/output transport) - Command:
npx
- Arguments:
-y
- Server name:
telerik_blazor_assistant
(depends on your preferences)
You also need to add your Telerik licence key as an env
parameter in the mcp.json
file. There are two options:
- (recommended) Use a
TELERIK_LICENSE_PATH
argument and point to your Telerik license file location. - Use a
TELERIK_LICENSE
argument and paste your Telerik license key. Make sure to update the license key when necessary.
Visual Studio
For detailed instructions, refer to Use MCP servers in Visual Studio.
Visual Studio 17.14 seems to impose the following requirements:
- Do not use hyphens (
-
) in the MCP server name in.mcp.json
. For example,"telerik_blazor_assistant"
and"telerikblazorassistant"
work, but"telerik-blazor-assistant"
does not.- The Copilot Chat window must be open and active when you open a solution.
To enable the Telerik MCP Server in a specific Blazor app, add a .mcp.json
file to the solution folder.
.mcp.json
{
"servers": {
"telerik_blazor_assistant": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@progress/telerik-blazor-mcp"],
"env": {
"TELERIK_LICENSE_PATH": "C:\\Users\\___\\AppData\\Roaming\\Telerik\\telerik-license.txt"
}
}
}
}
To enable global automatic discovery of the Telerik MCP Server in Visual Studio, add the above .mcp.json
file to your user directory (%USERPROFILE%
), for example, C:\Users\____\.mcp.json
.
Once the Telerik MCP server is added, make sure that the
telerik_blazor_assistant
tool is enabled (checked) in the Copilot Chart window's tool selection dropdown. The Telerik MCP server may get disabled and you may see "🔧10/11" in the selected tools dropdown when starting a new chat, changing threads, or relaunching Visual Studio. This is a known issue for MCPs that is being investigated.
VS Code
For detailed instructions, refer to Use MCP servers in VS Code.
To enable the Telerik MCP Server in a specific workspace or Blazor app, add a .vscode
folder with an mcp.json
file at the root of the workspace:
.vscode/mcp.json at the workspace root
{
"servers": {
"telerik_blazor_assistant": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@progress/telerik-blazor-mcp"],
"env": {
"TELERIK_LICENSE_PATH": "C:\\Users\\___\\AppData\\Roaming\\Telerik\\telerik-license.txt"
}
}
}
}
To add the Telerik MCP Server globally for VS Code, edit the VS Code settings.json
file:
VS Code settings.json
{
// ...
"chat.mcp.discovery.enabled": true,
"mcp": {
"servers": {
"telerik_blazor_assistant": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@progress/telerik-blazor-mcp"],
"env": {
"TELERIK_LICENSE_PATH": "C:\\Users\\___\\AppData\\Roaming\\Telerik\\telerik-license.txt"
}
}
}
}
}
Cursor
For detailed instructions, refer to Model Context Protocol.
To enable the Telerik MCP Server in a specific workspace or Blazor app, add a .cursor
folder with an mcp.json
file at the root of the workspace.
.cursor/mcp.json
{
"mcpServers": {
"telerik_blazor_assistant": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@progress/telerik-blazor-mcp"],
"env": {
"TELERIK_LICENSE_PATH": "C:\\Users\\___\\AppData\\Roaming\\Telerik\\telerik-license.txt"
}
}
}
}
To add the Telerik MCP Server globally for Cursor, add a .cursor
folder with the above mcp.json
file in your user folder.
Usage
To use the Telerik MCP Server:
-
Start your prompt with one of the following:
telerik
/telerik
@telerik
telerikblazor
/telerikblazor
@telerikblazor
-
Confirm that the Telerik MCP server is used, because this doesn't happen deterministically. Look for a statement in the output, which is similar to:
Running telerik_blazor_assistant
(in VS Code)Calling MCP tool telerik_blazor_assistant
(in Cursor)
If the Telerik MCP server is not used even though it's installed and enabled, then try rephrasing your prompt and use another trigger syntax from the list in step 1.
-
Grant the Telerik tool permission to run for this session, workspace, or always.
-
If you want to prompt for information or code that are not related to your previous prompts, it is a good practice to start a new session in a new chat window, so that the context is not polluted by irrelevant old information.
To increase the probability of the Telerik MVC Server being used, or to call it without the need to mention "telerik" explicitly, add custom instructions to your AI-powered tool. Here are examples for GitHub Copilot and Cursor.
Sample Prompts
The following list describes how your prompts may look like:
- "Telerik Generate a Blazor Grid with sorting and paging enabled. Bind the Grid to a Person model and provide dummy data."
- "Telerik Generate a ComboBox for Blazor that shows a list of products. Create a Product class and generate sample data."
- "Telerik Show me sample code for a Blazor Grid with virtual scrolling for the rows and columns."
Number of Requests
A Telerik Subscription license is recommended in order to use the Telerik Blazor AI Coding Assistant without restrictions. Perpetual license holders and trial users can make a limited number of requests per year.
Connect to Local AI Model
You can use the Telerik Blazor MCP server with local large language models (LLM). For example, run your local model through Ollama and use a third-party package such as MCP-LLM Bridge to connect the model to the Telerik MCP server. This will allow you to use the Telerik AI Coding Assistant without a cloud-based AI model.