Turning an API into relational tables sounds like a detour until you notice the effect: list the tables, read the columns, then write one query — a loop a model handles well and REST pagination it does not. Two things to plan for: the JDBC driver is a separate licensed install with its own trial, and this build is read-only, so anything that changes data needs CData's commercial server.
CData's read-only MCP server for eBay. The CData JDBC driver models eBay as relational tables, and this Java server puts a tool interface in front of that driver, so an assistant can discover the schema and run a SELECT rather than calling eBay's API directly.
- A tool that retrieves the tables available in the data source, returned as CSV with a header row
- A tool that retrieves the columns of a named table, so the model learns the schema before querying
- A tool that executes a SQL SELECT query
- Tool names carry the prefix you set in the .prp file, so several CData servers can coexist in one client
- Read-only by design — for write, update and delete, CData points at its own commercial MCP server
Java, plus Maven to build: `mvn clean install` produces CDataMCP-jar-with-dependencies.jar. Separately install and license the CData JDBC Driver for eBay, running `java -jar cdata.jdbc.ebay.jar --license` with your name, email and either TRIAL or a licence key. Then write a .prp file naming Prefix, ServerName, ServerVersion, DriverPath, DriverClass (cdata.jdbc.ebay.eBayDriver), the JdbcUrl from the connection-string utility, and optionally a table list. Launch as `java -jar` with the jar and the .prp path. It speaks stdio, so client and server must be on the same machine.
Build from source — clone the repository and build it, then deploy it and point your client at the endpoint