memgraph/memgraph is an in-memory graph database. A graph database is a storage structure that represents relationships with vertices and edges instead of the rows and columns of a table. On GitHub the primary language is C++, and the README states it is built in C/C++. The repository description presents the project for real-time graph analytics and AI context search. The repository was created on September 21, 2020, and the latest public release is v3.13.2.
It is a memory-centric graph DB that handles graph traversal together with vector and text indexes in a single query. Because text and vector similarity search and graph traversal are provided in one query layer, it is used for memory features that store and retrieve the context an AI agent has passed through, for agentic AI, and for graph-based retrieval-augmented generation (GraphRAG). It describes real-time graph analytics use cases for fraud detection, network analysis, and infrastructure monitoring. The query language is compatible with Neo4j's Cypher, and it claims ACID compliance and high availability. The repository has 4623 GitHub stars.
However, the GitHub metadata lists the license as NOASSERTION, so the license type cannot be confirmed. No changelog or performance measurements are given for v3.13.2. It is too early to judge whether the feature scope described in the README can be applied directly in real use.