Treating a Dropbox account as tables changes what you can ask: not just 'find this file' but 'what is in this folder tree, sorted by size and last modified'. Restricting the query tool to SELECT is what makes it safe to point at an account holding real work — the assistant can read but cannot move or delete. The price is a licensed JDBC driver, a Maven build and one browser authentication before the first question.
A local, read-only MCP server that exposes Dropbox as a relational SQL model. It wraps CData's JDBC driver for Dropbox, which presents the account's files, folders and metadata as tables, so an assistant can discover the schema and answer questions with live SELECT queries.
- The tables available in Dropbox, returned as CSV with a header row — `get_tables`
- The columns on a chosen table, so a query targets real column names — `get_columns`
- A SQL SELECT executed against live Dropbox data — `run_query`
- Tool names prefixed with the server name from your properties file, so several CData servers can coexist in one client
- A `Tables` property that limits which tables are exposed at all
A Dropbox account, plus CData's JDBC driver for Dropbox downloaded and licensed separately — `java -jar cdata.jdbc.dropbox.jar --license` takes your name, email and TRIAL or a license key. Java and Maven: `mvn clean install` produces the runnable JAR. Then a `.prp` file with the prefix, server name and version, the driver JAR path, the driver class `cdata.jdbc.dropbox.DropboxDriver`, and the JDBC connection string from the driver's connection-string utility — since Dropbox uses OAuth, that utility opens a browser to authenticate. The client entry runs java with the JAR and the `.prp` path. Communication is over stdio, so client and server must be on the same machine. MIT licensed.
Build from source — clone the repository and build it, then point your client at the binary