This connects Claude directly to an ArcadeDB instance, letting you query a multi-model database that supports SQL, Cypher, Gremlin, and GraphQL from the same conversation. You get access to graph, document, key/value, time series, and vector embedding models through their native query languages. The server exposes read and write operations across all these interfaces, so you can traverse graphs with Cypher, run SQL analytics, or search vectors without switching tools. Reach for this when you need Claude to work with complex relational data, knowledge graphs, or hybrid workloads where different query paradigms make sense for different questions. It's overkill for simple lookups but valuable when your data model demands flexibility.
claude mcp add --transport stdio arcadedata-arcadedb uvx arcadedb