memgraph
NoSQL & DocumentA high‑performance, in‑memory graph database that supports real‑time AI context retrieval and analytics with sub‑millisecond traversals.
Features
- Built‑in vector, text and geospatial indexes for hybrid similarity + graph queries
- 40+ native graph algorithms (PageRank, community detection, GNN link prediction, embeddings) in C++, Python and CUDA
- Atomic GraphRAG pipelines expressed as a single Cypher query
- In‑memory C/C++ engine delivering sub‑millisecond multi‑hop traversals
- Enterprise features: high availability, fine‑grained access control, encryption, backup/restore
Recent releases
View all 5 releases →
Config change
v3.10.0
Breaking risk
Auth
RBAC
Breaking upgrade
Init‑file rejection + DENY permissions
v3.9.0
Breaking risk
Breaking changes
- HA coordinator data directories must be cleaned and coordinators restarted; rerun setup queries on new coordinator leader
Notable features
- Vector index functions in WITH CONFIG
- Global server parameters
- Async logging support
v3.8.1
Bug fix
Fixed data instance deadlock that prevented startup under specific timing conditions.
v3.8.0
Breaking risk
Breaking changes
- Cannot switch to analytical mode if constraints exist
- Removed deprecated flags: --auth-module-executable, --bolt_session_inactivity_timeout, --storage_parallel_index_recovery
- New auth module message ID protocol required
Notable features
- Concurrent edge creation on supernodes
- Parallel Cypher execution (Enterprise)
- LOAD JSONL and LOAD PARQUET from HTTP/FTP/S3
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
About
Stars
4,099
Forks
231
Languages
C++
Python
Cypher
Install & Platforms
Install via
docker
brew
apt
dnf
pacman
zypper
binary
Platforms
linux
macos
windows
