Memgraph

Memgraph rounds out this category with a third specialist: an in-memory graph database rather than a relational store or a vector index, built for sub-millisecond traversals across connected data — fraud detection, network analysis, knowledge graphs and, increasingly, GraphRAG systems that give an AI agent structured context instead of only vector similarity.

The Community Edition is free and open source, with ACID transactions, replication and the Cypher query language, and Memgraph Cloud offers managed instances from 1GB up to 32GB of RAM on AWS, with automatic backups.

Memgraph Ltd has been registered in England since 23 May 2016, which puts it under UK jurisdiction rather than inside the EEA, the same position as SuiteCRM elsewhere in this directory.

Pricing scales with memory capacity rather than per query or per replica, and a 30-day free Enterprise trial is available alongside the always-free Community tier. Against Qdrant and Weaviate, Memgraph solves a different shape of problem — relationships between entities rather than semantic similarity between them — so a team doing both will likely end up running a graph database and a vector database side by side rather than choosing one.

What Memgraph does well

• Purpose-built in-memory graph database with sub-millisecond traversals
• Free, open-source Community Edition
• Cypher query language, ACID transactions
• Memory-based pricing, no per-query charges
• Registered in the UK since 2016

Where Memgraph falls short

• United Kingdom jurisdiction, outside the EEA
• Solves a different problem than the vector databases here — not interchangeable
• Memgraph Cloud region list not published on the vendor's own pages
• Smaller company than the established providers in this category

Standout feature. A graph database built for GraphRAG and real-time traversals — the connected-data counterpart to the vector databases sitting next to it.