Arcadedb

What ArcadeData/arcadedb MCP server does

ArcadeData/arcadedb exposes ArcadeDB through the Model Context Protocol. ArcadeDB is a transactional multi-model database that combines several data models in one system: graph, document, key-value, search, time-series, vector, and geospatial data. The database supports structured and unstructured records, ACID transactions, full-text indexes, geospatial queries, and security features.

The MCP server is relevant when an agent needs to inspect or work with ArcadeDB data using database-oriented operations. Its multi-model design can support applications that combine relationships, documents, embeddings, time-based measurements, or location-aware data instead of separating those concerns across different databases.

How it works

ArcadeDB includes MCP support as part of the database project. An MCP-compatible AI client connects to the database through that built-in server, allowing the client to use ArcadeDB capabilities in an agent workflow.

Queries can use several languages understood by ArcadeDB. SQL is available alongside Neo4j Cypher, Apache TinkerPop Gremlin, GraphQL, and MongoDB Query Language. The supported language depends on the data model and access pattern: graph workloads can use Cypher, Gremlin, or SQL; document workloads can use MongoDB queries or SQL; and other models have their corresponding database interfaces.

ArcadeDB also supports compatibility with several established database ecosystems. Its graph model is compatible with Cypher, Gremlin, and OrientDB SQL. The document model supports MongoDB drivers and queries, while the key-value model supports the Redis driver. Time-series access includes InfluxDB Line Protocol, Prometheus remote read and write, and PromQL.

Tools and capabilities

The database capabilities available behind the MCP interface include:

• Access to graph, document, key-value, search, time-series, vector, and geospatial data models
• SQL, Cypher, Gremlin, GraphQL, and MongoDB Query Language support
• Graph operations using linked records rather than relational joins
• Full-text indexing and geospatial querying
• Vector embedding storage and time-series data support
• Transactional operations through ArcadeDB’s ACID database engine
• Support for structured and unstructured data

ArcadeDB is implemented in Java and is designed to run on hardware ranging from a Raspberry Pi to cloud-hosted multi-server deployments. The database also includes graph features such as pathfinding, centrality, community detection, link prediction, and graph embeddings.