generalinformation-405551-help-resources-dev-express-documentation-mcp-server-configure-an-ai-powered-assistant.md
To meet growing demand for AI-assisted coding, we configured an MCP server that connects GitHub Copilot and other MCP-compatible AI tools directly to our documentation database.
Once you configure our MCP Server in your IDE, AI Assistants in Agent mode can help you with the following tasks:
When you submit DevExpress-related queries, add a reference to the configured MCP server at the end of your prompt.
Important
Always review AI-generated output: check for architectural problems, security vulnerabilities, and adherence to project standards. AI-generated output can vary greatly based on the prompt, AI model used, project code clarity, etc.
Supported IDEs:
Coding Assistants and AI Tools:
Since AI-related technology evolves rapidly and newer releases often include important fixes, we recommend that you always update to the latest version of your favorite IDE/plugins.
Tip
If you use the DevExpress Template Kit to create a new project in Visual Studio or VS Code, simply select the following option: Use DevExpress MCP. The Template Kit will set up an MCP Server automatically. Copilot will use this server to access DevExpress-specific guidance.
When you setup a documentation MCP server, the endpoint URL without parameters accesses documentation for the latest publicly available version. You can also specify a previous version as a parameter (no earlier than v24.2).
The MCP server offers two tools listed below. You can refer to these tools by name when you fine-tune your AI Assistant with custom instructions (setup examples available later in this article):
devexpress_docs_search - Performs semantic search through the documentation database and returns the top five matches for a user query.devexpress_docs_get_content - Allows agents to download complete help topics by URL.The server operates exclusively over the Streamable HTTP protocol. If you are familiar with this technology, add the following endpoints to your AI assistant’s configuration and give it a try:
{
"servers": {
"dxdocs": {
"url": "https://api.devexpress.com/mcp/docs",
"type": "http"
},
"dxdocs24_2": {
"url": "https://api.devexpress.com/mcp/docs?v=24.2",
"type": "http"
}
}
}
DevExpress MCP Server Endpoint: https://api.devexpress.com/mcp/docs.
Note
Specify this URL in a compliant MCP client that works with Streamable HTTP. This URL does not support direct access from a web browser and may return the following error if accessed manually: “405 Method Not Allowed”.
devexpress_docs_search and devexpress_docs_get_content.Use the following links for quick installation:
Note
If you experience issues with a one-click installation link, follow manual installation instructions listed in the next section.
Create .mcp.json in the %USERPROFILE% directory.
Create .mcp.json in the %USERPROFILE% directory.
Use Command Palette → “MCP: Add Server” or create <project>/.vscode/mcp.json.
Create <project>/.cursor/mcp.json or use Settings → Tools & Integrations → New MCP Server. Use "type": "sse" in Cursor configuration.
Go to Settings | Tools | AI Assistant | Model Context Protocol (MCP) and follow the wizard. Select the HTTP protocol and supply JSON configuration.
GitHub Copilot Plugin : Change chat mode to Agent and add tools. For additional information, see Extending GitHub Copilot Chat with MCP servers / JetBrains IDEs.
Configuration steps:
https://api.devexpress.com/mcp/docs.Learn more:
Run the following commands:
# Local configuration
claude mcp add --transport http dxdocs https://api.devexpress.com/mcp/docs
# Global configuration
claude mcp add --scope user --transport http dxdocs https://api.devexpress.com/mcp/docs
# Verify registered servers
claude mcp list
Learn more:
Note : Administrative privileges required.
Navigate to Repository → Settings → Copilot → Coding agent. Add JSON configuration in the “MCP configuration” section:
{
"dxdocs": {
"type": "http",
"url": "https://api.devexpress.com/mcp/docs",
"tools": ["*"]
}
}
Configuration steps:
https://api.devexpress.com/mcp/docs.Learn more:
Add to .gemini/settings.json:
{
"DevExpress MCP Server": {
"httpUrl": "https://api.devexpress.com/mcp/docs"
}
}
To obtain most relevant answers from DevExpress MCP Server:
Example Prompts:
To get the most from the DevExpress Documentation MCP server, define instructions that guide your AI assistant on how to effectively use DevExpress tools.
You can define custom instructions in all IDEs. Visual Studio and VS Code can use a predefined DevExpress prompt:
mcp.dxdocs.devexpress_docs_query_workflow
/mcp.dxdocs.devexpress_docs_query_workflowCreate or edit the file: <solution>/.github/copilot-instructions.md.
Sample instructions:
---
description: 'Answer questions about DevExpress UI Components and their API using the dxdocs server'
---
You are a .NET/JavaScript programmer and DevExpress product expert.
Your task is to answer questions about DevExpress components and their APIs using dxdocs MCP server tools.
When replying to **ANY** question about DevExpress components, use the dxdocs server to construct your answer.
## Workflow:
1. **Call devexpress_docs_search** to obtain help topics related to the user's question
2. **Call devexpress_docs_get_content** to fetch and read the most relevant help topics
3. **Reflect on the obtained content** and how it relates to the question
4. **Provide a comprehensive answer** based solely on retrieved information
## Constraints:
- **Use devexpress_docs_search only once** per question to avoid redundant queries
- **Answer questions based solely** on information obtained from MCP server tools
- If relevant code examples are available in documentation, **include those code examples**
- **Reference specific DevExpress controls and properties** mentioned in the docs
- If a user specifies a version (such as v24.2 or 24.2), invoke MCP server tools corresponding to that version (for example, "dxdocs24_2")
Learn more about GitHub Copilot custom instructions:
Navigate to Settings → Rules & Memories. Run the Add Rule command and specify custom instructions similar to the example above.
For additional information about Rules in Cursor IDE, refer to: Rules Documentation
Create a custom prompt as described in: Add and customize prompts
Claude Projects allow you to create dedicated workspaces with custom instructions and knowledge bases.
Claude Code uses CLAUDE.md files for custom instructions that load automatically at startup. You can create these files at different levels:
Project-level (shared with team, checked into version control):
./CLAUDE.md or ./.claude/CLAUDE.md in your project rootUser-level (personal settings across all projects):
~/.claude/CLAUDE.mdLocal project (personal project-specific, automatically git-ignored):
./.claude/CLAUDE.local.mdExample CLAUDE.md content:
# DevExpress Development Instructions
When working with DevExpress components:
- Always use the dxdocs MCP server to search DevExpress documentation.
- Reference specific DevExpress control names.
- Include technology stack (WinForms, WPF, Blazor, ASP.NET Core, VCL).
- Follow DevExpress best practices from official documentation.
## Common Commands
- Build: `dotnet build`
- Test: `dotnet test`
Alternative : Use the --append-system-prompt flag for ad-hoc instructions:
claude --append-system-prompt "Use dxdocs MCP server for all DevExpress questions"
For additional information, see Claude Code Memory Management.
Custom instructions can be set in Settings → Personalization → Custom instructions or included directly in your conversation prompts.