# Memgraph

A high-performance, in-memory graph database, compatible with Neo4j's Cypher, combining built-in vector and text indexes with graph traversal for GraphRAG, AI agent memory, and real-time graph analytics.

Memgraph is a high-performance, in-memory graph database, built in C/C++, that serves as the graph engine for GraphRAG pipelines, AI memory systems and agentic workflows, delivering sub-millisecond multi-hop traversals alongside semantic vector or text search. It combines built-in text, vector and geospatial indexes with full graph traversal in a single query layer, so a retrieval pipeline can run as one atomic database operation instead of being spread across several systems.

It includes the MAGE library of more than 40 graph algorithms in C++, Python and CUDA, including PageRank, community detection and GNN-based link prediction; an atomic GraphRAG capability expressing pivot search, expansion, ranking and prompt assembly as a single Cypher query; and real-time schema introspection for text-to-Cypher and agent integration. It is fully compatible with Neo4j's Cypher query language, is ACID-compliant and highly available, and can ingest streaming data from Kafka, Pulsar and RedPanda.

Website: https://memgraph.com