CData ADP Connector for Model Context Protocol (MCP)

:warning: This artifact establishes a strictly read-only MCP gateway. For comprehensive Create, Read, Update, Delete, and Action functionalities, alongside streamlined deployment, explore the CData MCP Server for ADP (Beta release).

Core Objective

This read-only MCP server permits Large Language Models (LLMs), such as Claude Desktop, to interrogate live data sourced from ADP, powered by the CData JDBC Driver for ADP.

The CData JDBC Driver translates ADP structures into familiar relational SQL models.

This server acts as a wrapper around that driver, exposing ADP assets via a straightforward MCP interface. Consequently, LLMs can retrieve current data simply by posing queries in natural language—eliminating the prerequisite for SQL scripting.

Implementation Procedure

1. Clone the source repository: bash git clone https://github.com/cdatasoftware/adp-mcp-server-by-cdata.git cd adp-mcp-server-by-cdata

2. Compile the server artifact: bash mvn clean install

   The resulting file is CDataMCP-jar-with-dependencies.jar 3. Acquire and install the necessary CData JDBC Driver for ADP: https://www.cdata.com/drivers/adp/download/jdbc 4. Activate the CData JDBC Driver license: * Navigate to the installed driver's lib directory (e.g., Windows: C:\Program Files\CData\CData JDBC Driver for ADP\ or Mac/Linux: /Applications/CData JDBC Driver for ADP/) * Execute the command: java -jar cdata.jdbc.adp.jar --license * Input your user details and the license identifier ("TRIAL" or your official key). 5. Define the data source connection parameters: * Launch the connection utility: java -jar cdata.jdbc.adp.jar

   • Configure the connection string (e.g., using Salesforce as a sample target) and validate with "Test Connection".

     Note: OAuth authentication necessitates interaction via a web browser.

   • Upon validation success, capture the generated connection string.
   • Construct a configuration file (e.g., adp.prp) for the JDBC linkage, specifying:
   • Prefix: Namespace for exposed tools.
   • ServerName: Identifier for this MCP instance.
   • ServerVersion: Version reporting.
   • DriverPath: Absolute location of the JDBC driver JAR.
   • DriverClass: Fully qualified class name of the JDBC driver (e.g., cdata.jdbc.adp.ADPDriver).
   • JdbcUrl: The connection string obtained in step 5.
   • Tables: Specify desired tables, or leave empty for full schema exposure. env Prefix=adp ServerName=CDataADP ServerVersion=1.0 DriverPath=ABSOLUTE/PATH/TO/cdata.jdbc.adp.jar DriverClass=cdata.jdbc.adp.ADPDriver JdbcUrl=jdbc:adp:InitiateOAuth=GETANDREFRESH; Tables=

Integration with AI Clients (e.g., Claude Desktop)

1. Prepare the client configuration file (claude_desktop_config.json) to integrate the new MCP service. Append the server definition to the mcpServers object:

   Windows Client Configuration Snippet

   { "mcpServers": { "{classname_dash}": { "command": "C:\PATH\TO\java.exe", "args": [ "-jar", "C:\PATH\TO\CDataMCP-jar-with-dependencies.jar", "C:\PATH\TO\adp.prp" ] }, ... } }

   Linux/Mac Client Configuration Snippet

   { "mcpServers": { "{classname_dash}": { "command": "/PATH/TO/java", "args": [ "-jar", "/PATH/TO/CDataMCP-jar-with-dependencies.jar", "/PATH/TO/adp.prp" ] }, ... } }

   If necessary, relocate the finalized configuration file to the client's designated user directory (e.g., Windows: %APPDATA%\Claude\, Mac/Linux: /Users/{user}/Library/Application\ Support/Claude/). 2. Relaunch or refresh your consuming application to register the newly added MCP endpoint.

Crucial Note: A full application shutdown and restart (not merely closing the window) may be required for the MCP service definitions to take effect.

Standalone Server Execution

To initiate the MCP service independently: bash java -jar /ABSOLUTE/PATH/TO/CDataMCP-jar-with-dependencies.jar /ABSOLUTE/PATH/TO/adp.prp

Constraint: Communication relies exclusively on standard I/O (stdio), limiting operation to clients running on the identical host machine.

Interaction Model

Upon successful MCP configuration, the AI agent gains the capability to issue read, write, update, and delete operations against the underlying ADP data via its internal toolset. Explicit tool invocation is rarely needed; direct natural language requests suffice. Examples of effective queries include: * "Determine the statistical relationship between closed/won opportunities and the associated account sector." * "Provide a count of active support cases within the 'SUPPORT' functional area." * "List my scheduled calendar entries for the current day."

Available Toolset & Documentation

The reference {servername} corresponds to the identifier defined in the configuration file (e.g., {classname_dash}). * {servername}_get_tables: Fetches the schema's tabular inventory. Use {servername}_get_columns subsequently to detail fields per table. Output is formatted as CSV, starting with headers. * {servername}_get_columns: Retrieves the attribute list for a specified table. Always predate this with {servername}_get_tables. Output is formatted as CSV, starting with headers. * {servername}_run_query: Executes a provided SQL SELECT statement against the data model.

JSON-RPC Command Schemas

For scenarios involving direct scripting or programmatic interaction (bypassing the AI interface), requests must adhere to the JSON-RPC 2.0 standard, targeting the available methods:

Invoking Schema Table Retrieval (adp_get_tables)

{ "jsonrpc": "2.0", "id": 1, "method": "tools/call", "params": { "name": "adp_get_tables", "arguments": {} } }

Invoking Field Retrieval (adp_get_columns)

{ "jsonrpc": "2.0", "id": 2, "method": "tools/call", "params": { "name": "adp_get_columns", "arguments": { "table": "AccountRecord" } } }

Executing a Data Retrieval Command (adp_run_query)

{ "jsonrpc": "2.0", "id": 3, "method": "tools/call", "params": { "name": "adp_run_query", "arguments": { "sql": "SELECT Name, Status FROM [Tickets] WHERE Priority = 'High'" } } }

Troubleshooting Guide

1. Visibility Issue: If the CData MCP Server is not appearing in the AI client, ensure the client process has been completely terminated (check Task Manager/Activity Monitor).
2. Data Retrieval Failure: Verify the connection parameters are correct. Use the built-in connection string utility to generate and accurately transfer the string into the .prp file.
3. Connectivity Problems: For issues connecting to the target ADP system, engage the CData Support Team.
4. MCP Server Feedback: For general server usage concerns or feature suggestions, participate in the CData Community.

Licensing Details

This MCP server software is distributed under the terms of the MIT License, granting rights to use, alter, and redistribute, subject to the license conditions outlined in the repository's LICENSE file.

Comprehensive Supported Connectors

[Table Listing All Supported Data Sources (Access, Salesforce, SAP, Snowflake, etc.) - See Original Source for Full Table]

(Excerpt for reference) | Data Source | Data Source | Data Source | Data Source | | :--- | :--- | :--- | :--- | | Access | Act CRM | Act-On | Active Directory | | ADP | Airtable | AlloyDB | Amazon Athena | | ... | ... | ... | ... | | Xero | XML | YouTube Analytics | Zendesk |