OpenLink MCP JDBC Server

You can run an MCP server for JDBC to securely access and query relational databases from MCP clients. This server connects via JDBC to backends such as Virtuoso and other JDBC-enabled DBMSs, exposing a set of MCP endpoints to discover schemas, list tables, describe structures, run queries, and even execute Virtuoso-specific SPARQL/SPASQL operations.

How to use

Use an MCP client to connect to the JDBC MCP server and call the exposed tools to explore your database and run queries. The server supports common actions like listing schemas and tables, describing a table’s columns, filtering tables by name, and executing SQL queries in JSONL or Markdown formats. Virtuoso-specific features such as SPASQL and SPARQL queries, as well as the Virtuoso AI assistant, are also available.

How to install

Prerequisites: Ensure you have Java 21 or higher installed on your system.

Clone the repository and navigate into the project directory.

Run the Java MCP server using the provided runner JAR. You will typically start the server with a command like this:

java -jar /path/to/mcp-jdbc-server/MCPServer-1.0.0-runner.jar

Additional configuration and runtime notes

The server reads JDBC connection details from environment variables. The following variables are used in examples and should be set when starting the server or provided to MCP client configurations.

{
  "jdbc.url": "jdbc:virtuoso://localhost:1111",
  "jdbc.user": "username",
  "jdbc.password": "password",
  "jdbc.api_key": "sk-xxx"
}

Claude Desktop specific configuration for Virtuoso JDBC

If you are using Claude Desktop with Virtuoso and its JDBC driver, add the following configuration to your Claude Desktop setup so the MCP server runs as a local Java process.

{
  "mcpServers": {
    "my_database": {
      "command": "java",
      "args": ["-jar", "/path/to/mcp-jdbc-server/MCPServer-1.0.0-runner.jar"],
      "env": {
        "jdbc.url": "jdbc:virtuoso://localhost:1111",
        "jdbc.user": "username",
        "jdbc.password": "password",
        "jdbc.api_key": "sk-xxx"
      }
    }
  }
}

Claude Desktop with additional JDBC drivers

If you use Claude Desktop with another JDBC driver or multiple drivers, add an extended classpath configuration so all required drivers are available to the MCP server.

{
  "jdbc": {
    "command": "java",
    "args": [
      "-cp",
      "/path/to/mcp-jdbc-server/MCPServer-1.0.0-runner.jar:/path/to/jdbc_driver1.jar:/path/to/jdbc_driverN.jar",
      "io.quarkus.runner.GeneratedMain"
    ],
    "env": {
      "jdbc.url": "jdbc:virtuoso://localhost:1111",
      "jdbc.user": "dba",
      "jdbc.password": "dba"
    }
  }
}

Troubleshooting with MCP Inspector

You can verify and troubleshoot MCP server interactions using the MCP Inspector. Install the inspector, then start it with the Java runner to inspect communications between the MCP client and the server.

npm install -g @modelcontextprotocol/inspector

Run the inspector against the MCP server

To connect the inspector to the local MCP server, start the inspector with the Java classpath including the MCP runner JAR and any JDBC driver JARs you need.

npx @modelcontextprotocol/inspector java -cp MCPServer-1.0.0-runner.jar:/path/to/driver1.jar:/path/to/driverN.jar io.quarkus.runner.GeneratedMain

Use example for querying

After the inspector starts, access the URL it returns and use the available MCP operations to run queries such as listing schemas, fetching tables, describing a table, or executing SQL queries in JSONL or Markdown formats.

Available tools

jdbc_get_schemas

Retrieve and return a list of all schema names from the connected database.

jdbc_get_tables

Retrieve and return a list of tables for a specified schema or the default schema.

jdbc_filter_table_names

Filter and return information about tables whose names contain a substring.

jdbc_describe_table

Return detailed information about the columns of a specific table.

jdbc_query_database

Execute a standard SQL query and return results in JSON format.

jdbc_query_database_md

Execute a standard SQL query and return results formatted as a Markdown table.

jdbc_query_database_jsonl

Execute a standard SQL query and return results in JSONL format.

jdbc_spasql_query

Execute a SPASQL query and return results (Virtuoso-specific).

jdbc_sparql_query

Execute a SPARQL query and return results (Virtuoso-specific).

jdbc_virtuoso_support_ai

Interact with Virtuoso Support AI to run prompt-based queries.