Claude Code is a developer-focused environment with built-in AI features. It supports natural language interaction with structured data through CData's Model Context Protocol (MCP).

This topic explains how to configure the CData Code Assist MCP to work with Claude Code. CData Code Assist MCP for SAPSuccessFactorsLMS allows Claude Code to securely query and describe live data using plain language prompts.

Prerequisites

Before you begin, make sure the following are installed and configured:

• The CData Code Assist MCP for SAPSuccessFactorsLMS is installed.
• A connection has been created and tested in the CData Code Assist MCP Configuration Tool.
• Claude Code CLI is installed on your machine. For more details, see Claude Code overview.

Configuring CData Code Assist MCP

The recommended way to configure an MCP server for Claude Code is to copy the MCP server configuration details directly from the CData Code Assist MCP Configuration Tool UI and paste them into the appropriate configuration file for your selected scope. This ensures that connection names, paths, and arguments are accurate and eliminates the need to manually assemble command-line parameters.

Alternatively, Claude Code supports registering MCP servers through its command-line tool. The location where configuration is stored depends on the selected scope.

Configuration Scopes

Claude Code supports three types of configuration scopes for MCP servers:

Scope	Location	Shareable	Description
User	C:\Users\[YourUsername]\.claude.json	No	Applies to all Claude Code projects on the machine. This configuration is personal and not intended for sharing across machines or repositories.
Project	<project-root>/.mcp.json	Yes	Applies only to the current project. This configuration can be committed to version control and shared with team members.
Local	C:\Users\[YourUsername]\.claude.json	No	Applies only to the current project. This configuration is not intended for sharing. To prevent this configuration from being shared in the repository, add it to the project's .gitignore file.

Steps

1. If you are copying configuration from the UI, paste the JSON content into the appropriate configuration file for your selected scope.
2. If you prefer to register the server manually, run the claude mcp add command with your desired configuration scope and provide the path to the Java executable and JAR file.
3. Confirm that the server name matches the MCP connection name configured in the CData Code Assist MCP Configuration Tool.
4. Restart Claude Code or open a new Claude Code session to load the updated configuration.

Example Command-Line Registration

The following examples show how an MCP server configuration appears when registered for Claude Code. They illustrate the relationship between the command-line registration and the resulting JSON configuration file.

Update the values only if you are creating the configuration manually:

• Connection name: The server name (for example, MyDataSource) must match the connection name defined in the CData Code Assist MCP Configuration Tool exactly and is case-sensitive.
• Java path: The Java executable path must point to the Java runtime bundled with the MCP server.
• JAR path: The JAR file path must reference the source-specific MCP server JAR file.

When you register an MCP server using the Claude Code CLI, Claude Code writes the configuration to the appropriate JSON file for the selected scope. The following example shows a user-scoped registration and the corresponding configuration that appears in the .claude.json file.

Note: The .claude.json file contains all of your Claude Code user settings, not just MCP servers.

claude mcp add --transport stdio --scope user MySAPSuccessFactorsLMSDataSource -- "C:\Program Files\CData\CData Code Assist MCP for SAPSuccessFactorsLMS 2025\jre\bin\java.exe" -Dfile.encoding=UTF-8 -jar "C:/Program Files/CData/CData Code Assist MCP for SAPSuccessFactorsLMS 2025/lib/cdata.mcp.sapsuccessfactorslms.jar" MySAPSuccessFactorsLMSDataSource

A typical .claude.json entry for a registered MCP Server might look like the following:

"mcpServers": {
    "MySAPSuccessFactorsLMSDataSource": {
      "type": "stdio",
      "command": "C:\\Program Files\\CData\\CData Code Assist MCP for SAPSuccessFactorsLMS 2025\\jre\\bin\\java.exe",
      "args": [
        "-Dfile.encoding=UTF-8",
        "-jar",
        "C:/Program Files/CData/CData Code Assist MCP for SAPSuccessFactorsLMS 2025/lib/cdata.mcp.sapsuccessfactorslms.jar",
        "MySAPSuccessFactorsLMSDataSource"
      ],
      "env": {}
    }
  }

Verifying Configuration

After registration, verify that the MCP server is recognized.

Steps

1. Open a terminal and run claude mcp list.
2. Check that your server name appears with a "Connected" status in the output.
3. Start Claude Code by running claude inside your project.
4. Inside a Claude Code session, type /mcp to view active servers.

Managing Multiple MCP Servers

Claude Code supports registering multiple MCP servers across user, project, or local scopes. Servers can be added either by copying configuration from the CData Code Assist MCP Configuration Tool UI or by using the claude mcp add command.

The following example registers a project-scoped MCP server named MyDataSource:

claude mcp add --transport stdio --scope project MyDataSource -- "C:\Path\To\Java\bin\java.exe" -Dfile.encoding=UTF-8 -jar "C:/Path/To/MCPServer/lib/cdata.mcp.example.jar" MyDataSource

Troubleshooting

The server is not recognized

• Restart your terminal and Claude Code session to reload configuration changes. You may also need to initiate a Claude Code session before running claude mcp list. This step grants approval for Claude Code to use the newly added MCP server.
• Double-check that the server name passed to the claude mcp add command matches the connection name in the CData Code Assist MCP Configuration Tool. Names are case-sensitive.
• Run claude mcp list to verify the server is properly registered and appears with a valid status.

Connection issues

• Verify that the Java executable path is correct. It should point to the java.exe bundled with your MCP server (typically in a jre/bin subdirectory).
• Confirm that the JAR file path is valid and points to the correct source-specific JAR file inside the MCP server's lib folder (for example: cdata.mcp.sapsuccessfactorslms.jar).
• Check the MCP server configuration in the CData Code Assist MCP Configuration Tool:
  • Ensure the connection you're referencing is listed and active.
  • Click Save & Test to verify successful connectivity to the data source.

File path issues

• For terminal commands (like claude mcp add), use Windows-style paths with single backslashes (C:\Program Files\...) or forward slashes (C:/Program Files/...). Both are valid in most command-line tools.
• For JSON files (such as .claude.json or .mcp.json), use either double backslashes (C:\\Program Files\\...) or forward slashes (C:/Program Files/...) to ensure the file parses correctly.
• If the MCP server does not start, verify that the full file path to both the Java executable and the JAR file is correct and matches the version installed.

Authentication errors

• Launch the CData Code Assist MCP Configuration Tool and select the relevant connection.
• Use the Save & Test button to confirm that credentials are valid.
• If the test fails, re-enter or update your username, password, or token as needed.
• Save the connection after updating credentials, then restart Claude Code to reload the updated authentication context.