azure-devops -by-cdata

This MCP server provides a read-only interface to Azure DevOps data via the CData JDBC Driver. It wraps the driver in a lightweight MCP server so large language models (LLMs) can query live Azure DevOps data without writing SQL or managing direct connections. After starting the MCP server, you can connect using Claude Desktop or any MCP client that supports the standard MCP JSON-RPC tool APIs. The server exposes tools such as azure_devops_get_tables, azure_devops_get_columns, and azure_devops_run_query to discover data and run read-only queries. Use the built-in tools by issuing high-level natural language requests or by calling the JSON-RPC endpoints to enumerate tables, fetch column metadata, or run SELECT queries against the underlying Azure DevOps data model.

In Claude Desktop, add a new MCP server configuration pointing to the Java-based MCP process (the jar and the .prp file). The tools follow a naming convention where {servername} is the server identifier (e.g., azure_devops). Examples include azure_devops_get_tables, azure_devops_get_columns, and azure_devops_run_query. When using Claude or any compatible client, you can invoke these tools directly via JSON-RPC, for example asking the client to list tables or to execute a SELECT against a specific table. The server operates in stdio mode, meaning it communicates with clients on the same host process or session, so ensure the client and server are on the same machine as described in the setup notes.

Prerequisites

• Java Development Kit (JDK) installed and JAVA_HOME configured.
• Maven installed for building the MCP server.
• Access to the CData JDBC Driver for Azure DevOps (download and license as described in the README).

Installation steps

1. Clone the repository and navigate into it:

git clone https://github.com/cdatasoftware/azure-devops-mcp-server-by-cdata.git
cd azure-devops-mcp-server-by-cdata

2. Build the MCP server JAR using Maven:

mvn clean install

This produces a JAR such as CDataMCP-jar-with-dependencies.jar. 3) Download and install the CData JDBC Driver for Azure DevOps from: https://www.cdata.com/drivers/azuredevops/download/jdbc/ 4) License the JDBC Driver (as described in the README) and set up your azure-devops.prp file with connection details (Prefix, ServerName, ServerVersion, DriverPath, DriverClass, JdbcUrl, Tables). 5) Create or edit the Claude Desktop configuration file to include the new MCP server entry. For Windows, place the claude_desktop_config.json in the appropriate AppData path; for Linux/Mac, place it in your user library config path. 6) Run the MCP server:

java -jar /path/to/CDataMCP-jar-with-dependencies.jar /path/to/azure-devops.prp

Note: The server uses stdio, so it is intended to be run on the same machine as the client. 7) Start Claude Desktop (or your MCP client) and refresh to see the new server. Then invoke the available tools to query Azure DevOps data.

Tips and considerations:

• This MCP server is read-only by design. To enable full read/write capabilities against Azure DevOps, use the full MCP server variant described in the README (not this read-only version).
• The server communicates via stdio, so run both client and server on the same machine, or use a wrapper to connect from a local client session.
• The JSON-RPC tool names are based on the server name (azure_devops). Use azure_devops_get_tables, azure_devops_get_columns, and azure_devops_run_query to interact with data.
• If Claude Desktop does not immediately detect the MCP server, fully quit and re-open Claude Desktop after starting the server.
• Ensure the JDBC Driver is properly licensed and the connection string (JdbcUrl) is tested via the driver’s Connection String utility before creating the .prp file.