Skip to content
SPECIFICTION
Get listed

Memgraph

memgraph.com

Free plan WebWindowsmacOSLinuxAPISelf-hosted
Visit memgraph.com
Spec score 8.6 #6 of 39 in Graph Databases

What it is

Memgraph is an in-memory graph database platform for graph computations on streaming data. It supports Cypher and lists Kafka, Pulsar, and Redpanda as streaming connectors. Its free Community Edition includes ACID transactions, on-disk persistence, high-availability replication, and vector search. MAGE adds more than 60 graph algorithms, with integrations for NetworkX, igraph, and cuGraph. Memgraph Lab provides a visual interface for querying, managing, importing, and visualizing graph data; it also supports natural-language queries through GraphChat. Programmatic drivers are listed for Python, C/C++, Rust, Node.js, C#, Go, Haskell, Java, JavaScript, PHP, and Ruby. Enterprise features include role- and label-based access controls, SSO, LDAP/PAM authentication, TLS encryption in transit, audit logging, failover, replication, disaster recovery, no-downtime updates, and monitoring. Community support is available through Discord; Enterprise includes direct engineering support through a dedicated Slack channel. Memgraph Cloud is managed and hosted on AWS in six regions, with instances from 1 GB to 32 GB RAM. Community Edition is free forever. Enterprise plans have unlisted prices that scale by memory capacity or graph data capacity, depending on the edition, and the Enterprise trial lasts 30 days.

Spec sheet

What it does
Memgraph is an in-memory C++ graph database for real-time AI context, GraphRAG, agent memory and graph analytics.Read 15 Oct 2026 · memgraph.com
Product
Memgraph is an in-memory graph database and platform focused on graph computations on streaming data.Read 29 Sep 2026 · memgraph.com
Deployment
Memgraph supports Docker, native Linux packages, Kubernetes and cloud deployments on AWS, Azure and GCP.Read 15 Oct 2026 · memgraph.com
Trial
Memgraph offers a free 30-day Enterprise trial.Read 29 Sep 2026 · memgraph.com
Security
Enterprise features include role and label based access controls, SSO, LDAP/PAM authentication, TLS encryption in transit, and query audit logging.Read 29 Sep 2026 · memgraph.com
Support
Community Edition lists Discord support; Enterprise includes direct engineering support through a dedicated Slack channel.Read 29 Sep 2026 · memgraph.com
Headquarters
Sevenoaks, Kent, United KingdomRead 23 Sep 2026 · memgraph.com
Founded
2016Read 23 Sep 2026 · memgraph.com
Agentic AI
Reasoning graphs expose agent action paths and make traversed evidence inspectable.Read 15 Oct 2026 · memgraph.com
AI integrations
The documentation provides LangChain and LangGraph tooling, including a Memgraph toolkit and MemgraphQAChain.Read 15 Oct 2026 · memgraph.com
AI memory
Memgraph models semantic, episodic and procedural long-term memory as a queryable graph.Read 15 Oct 2026 · memgraph.com
Algorithms
MAGE provides 60+ graph algorithms, with integrations for NetworkX, igraph, and cuGraph.Read 29 Sep 2026 · memgraph.com
Availability
Enterprise includes automatic failover, synchronous and asynchronous replication, disaster recovery, and no-downtime updates.Read 29 Sep 2026 · memgraph.com
Built-in algorithms
Core traversal procedures include breadth-first search, depth-first search, weighted shortest path and all shortest paths.Read 15 Oct 2026 · memgraph.com
Client libraries
Official documentation lists clients or guides for C#, C/C++, Go, GraphQL, Haskell, Java, JavaScript, Node.js, PHP, Python, Ruby, Rust and Swift.Read 15 Oct 2026 · memgraph.com
Cloud infrastructure
Memgraph Cloud is fully managed on AWS, offers six regions, and supports instances up to 32 GB RAM and eight CPU cores.Read 15 Oct 2026 · memgraph.com
Cloud limits
Memgraph Cloud lists instance sizes from 1 GB to 32 GB RAM and says the service is AWS hosted in six regions.Read 29 Sep 2026 · memgraph.com
Cloud security
Cloud connections use SSL encryption and Cloud instances include Enterprise role-based access control.Read 15 Oct 2026 · memgraph.com
Company headquarters and founding
LinkedIn lists Memgraph as headquartered in London, England, founded in 2016, privately held and employing 11–50 people.Read 15 Oct 2026 · linkedin.com
Company history
Memgraph’s about page says its founders celebrated the company’s sixth year in 2022.Read 29 Sep 2026 · memgraph.com
Company legal name
Memgraph’s product pages identify the company as Memgraph Ltd.Read 15 Oct 2026 · memgraph.com
Custom modules
Developers can extend Memgraph with query modules written in C, C++, Python or Rust, with optional NVIDIA cuGraph acceleration.Read 15 Oct 2026 · memgraph.com
Documentation depth
The docs cover installation, client libraries, migration from graph and SQL systems, query modules, deployment, configuration, troubleshooting and release notes, and expose a public MCP documentation server.Read 15 Oct 2026 · memgraph.com
Drivers
The capabilities page lists programmatic drivers for Python, C/C++, Rust, Node.js, C#, Go, Haskell, Java, JavaScript, PHP, and Ruby.Read 29 Sep 2026 · memgraph.com
Funding
Memgraph announced a $9.34 million round led by Microsoft’s M12 venture fund with Heavybit, In-Q-Tel, Counterview Capital, ID4 Ventures and Mundi Ventures participating.Read 15 Oct 2026 · memgraph.com
GraphRAG
The product combines vector or text search with multi-hop graph traversal so retrieval can follow connected entities rather than similarity alone.Read 15 Oct 2026 · memgraph.com
Independent licence assessment
Is It Really FOSS concludes that Memgraph is not FOSS because BSL 1.1 and the Memgraph Enterprise License do not provide FOSS rights, despite open-source marketing.Read 15 Oct 2026 · isitreallyfoss.com
Independent review limitation
Hysen Labs notes that benchmark speed claims are vendor statements until reproduced and flags the mixed licence model as an adoption cost.Read 15 Oct 2026 · hysenlabs.com
Installation platforms
The download page provides one-command installers for Linux/macOS and PowerShell installation for Windows.Read 15 Oct 2026 · memgraph.com
Lab
Memgraph Lab provides a visual interface to query, manage, import, and visualize graph data, and includes natural-language querying through GraphChat.Read 29 Sep 2026 · memgraph.com
Licence
The repository states that source files are variously licensed under Business Source License 1.1 and the Memgraph Enterprise License.Read 15 Oct 2026 · github.com
Limit: production reliability complaint
A Reddit user reported production-level bugs that made Memgraph unusable for their real-time application, while another user said it worked well without issues.Read 15 Oct 2026 · reddit.com
Limit: Python client dependencies
A G2 reviewer says the Python client pulls in many dependencies that may not all be useful.Read 15 Oct 2026 · g2.com
Limit: RAM-bound sizing
An independent architecture review says Memgraph’s in-memory design delivers fast hot-set reads but creates a capacity ceiling tied to available RAM and trails Neo4j in ecosystem maturity.Read 15 Oct 2026 · shirokoff.ca
MAGE
MAGE supplies more than 60 graph algorithms, including PageRank, Leiden community detection, node2vec and betweenness centrality.Read 15 Oct 2026 · memgraph.com
Monitoring
Enterprise monitoring includes Prometheus metrics and query and transaction latency tracking.Read 29 Sep 2026 · memgraph.com
Notable release change
Release notes state that version 2.15.0, dated February 28, 2024, strengthened password hashing by adding salts and migrating hashes after successful login.Read 15 Oct 2026 · memgraph.com
Other integrations
The integrations catalog includes NetworkX, igraph, cuGraph and Elasticsearch.Read 15 Oct 2026 · memgraph.com
Performance claims
The maker advertises sub-millisecond traversals, more than 1,000 transactions per second and graph sizes from 100 GB to 4 TB.Read 15 Oct 2026 · memgraph.com
Praise: documentation and community
A G2 reviewer describes the documentation as clear and accessible and the core team as active and helpful.Read 15 Oct 2026 · g2.com
Praise: performance and usability
G2 reviewers praise Memgraph for strong performance, ease of use, fast memory-efficient algorithms and simple Cypher-based analytics.Read 15 Oct 2026 · g2.com
Query language
Memgraph supports the Cypher query language.Read 29 Sep 2026 · memgraph.com
Rating on G2
5.0 out of 5 from 2 reviews.Read 15 Oct 2026 · g2.com
Research benchmark context
A 2026 arXiv evaluation benchmarked Memgraph alongside six other graph or graph-like engines and found no system universally fastest across workloads.Read 15 Oct 2026 · arxiv.org
Storage modes
The database documentation lists in-memory transactional, in-memory analytical and on-disk transactional modes.Read 15 Oct 2026 · memgraph.com
Streaming
The product lists Kafka, Pulsar, and Redpanda as streaming connectors.Read 29 Sep 2026 · memgraph.com
Streaming integrations
Native streaming connectors are listed for Kafka, Redpanda and Pulsar.Read 15 Oct 2026 · memgraph.com
Transactions
Community Edition includes ACID transactions with on-disk persistence and high-availability replication.Read 29 Sep 2026 · memgraph.com
Transactions and persistence
Memgraph states that it provides ACID transactions in memory while persisting writes to permanent storage.Read 15 Oct 2026 · memgraph.com
Use cases
The company describes enterprise uses including AI infrastructure, GraphRAG, AI memory, fraud detection, network analysis, supply chain dependencies, and infrastructure monitoring.Read 29 Sep 2026 · memgraph.com

Plans and prices

Community Edition Free free forever in-memory graph database · ACID persistence · high-availability replication · Cypher · Kafka/Pulsar/Redpanda connectors · vector search · MAGE · Discord support Read 15 Oct 2026 · memgraph.com
Enterprise trial Free 30 days full Enterprise feature evaluation Read 15 Oct 2026 · memgraph.com
Memgraph Cloud Free 14-day trial; paid usage after trial AWS hosted in 6 regions · 1–32 GB RAM instances · one 2 GB project during trial · maximum 3 paid projects Read 15 Oct 2026 · memgraph.com
Enterprise Edition Priced on request contact sales; memory-capacity pricing RBAC and label-based access · SSO · LDAP/PAM · multi-tenancy · automatic failover · disaster recovery · audit logging · Prometheus · Slack engineering support Read 15 Oct 2026 · memgraph.com
Enterprise Edition — AI Platform Priced on request Price scales with graph data capacity only. Unlimited vector indexes · expanding AI tooling · enterprise security and access controls · multi-tenancy · automatic failover · disaster recovery · audit logging · Prometheus monitoring · dedicated engineering support Read 29 Sep 2026 · memgraph.com
Enterprise Edition — Graph Analytics Priced on request Price scales with total memory capacity; graph data and vector indexes counted together. All-inclusive · enterprise security and access controls · SSO · multi-tenancy · automatic failover · disaster recovery · audit logging · Prometheus monitoring · dedicated engineering support Read 29 Sep 2026 · memgraph.com
OEM / SaaS licence Priced on request usage-based commercial agreement embedded on-premises or SaaS distribution · single- or multi-tenant · engine remains embedded Read 15 Oct 2026 · memgraph.com

Every Memgraph plan and price →

Strong points and limits

Strong points

  • Free Community Edition includes ACID transactions.
  • MAGE provides more than 60 graph algorithms.
  • Lab includes visual and natural-language querying.
  • Drivers span 12 listed programming languages.

Limits

  • Cloud instances range from 1 GB to 32 GB RAM.
  • Enterprise trial is limited to 30 days.

Who it suits

Memgraph suits developers and teams working with graph data, particularly where streaming connectors, graph algorithms, or vector search are useful. Its Community Edition offers a free route to self-managed use, while Enterprise and Cloud address teams seeking additional controls or managed hosting.

Memgraph in depth

Read the full write-up

Overview

Memgraph is an in-memory graph database for teams that need to query and compute over streaming data. It suits developers and data teams building graph workloads who want to start with an open-source edition and add managed hosting or enterprise controls when needed. Its strongest case is the combination of streaming connectors, graph algorithms and a capable free edition; the main caveat is that paid options use custom pricing.

Cypher queries, vector search and ACID transactions anchor the database. The Community Edition adds on-disk persistence and high-availability replication, making it a substantial self-hosted starting point rather than a stripped-down query demo. Memgraph’s stated use cases span fraud detection, network analysis, supply-chain dependencies, infrastructure monitoring and AI workloads such as GraphRAG and AI memory.

For a broader comparison, see Graph Databases.

Key features

Streaming and graph algorithms

Memgraph names Kafka, Pulsar and Redpanda as streaming connectors. That makes it a natural candidate when graph computations need to work with streaming data rather than only a static graph. MAGE provides more than 60 graph algorithms and integrates with NetworkX, igraph and cuGraph, giving teams options to work across familiar graph and numerical-computing ecosystems.

Querying and visual work

Cypher is the supported query language. Memgraph Lab provides a visual interface for querying, managing, importing and visualizing graph data, with natural-language queries through GraphChat. That is useful for teams that want a visual counterpart to programmatic access, though the interface does not replace the need to understand the underlying graph model and queries.

Programmatic drivers cover Python, C/C++, Rust, Node.js, C#, Go, Haskell, Java, JavaScript, PHP and Ruby. The breadth makes it practical to connect Memgraph to varied application stacks. Vector search is also included, relevant to graph workloads that combine relationships with vector indexes.

Pricing

Memgraph uses a freemium model. Community Edition is free and is free forever, open source, and includes the full in-memory graph database, ACID transactions with on-disk persistence, high-availability replication, Cypher, streaming connectors, vector search, MAGE algorithms and Discord community support. It is the clearest fit for teams able to operate their own deployment and comfortable relying on community support.

Enterprise Edition — Graph Analytics has custom pricing, billed according to total memory capacity; graph data and vector indexes count together. It adds enterprise security and access controls, SSO, multi-tenancy, automatic failover, disaster recovery, audit logging, Prometheus monitoring and dedicated engineering support. Choose it when production governance and resilience matter more than keeping the deployment on the free tier.

Enterprise Edition — AI Platform also has custom pricing, billed by graph data capacity only. It includes unlimited vector indexes and expanding AI tooling alongside enterprise security and access controls, multi-tenancy, automatic failover, disaster recovery, audit logging, Prometheus monitoring and dedicated engineering support. This capacity basis may suit workloads where vector indexes would otherwise weigh heavily in the Graph Analytics pricing calculation.

Memgraph Cloud has custom pricing and is fully managed, with AWS hosting in six regions, instances from 1 GB to 32 GB RAM, automatic updates and backups, and community support. It trades the self-management of Community Edition for managed operations, but does not include the dedicated engineering support described for Enterprise. A free 30-day Enterprise trial offers a time-limited way to assess the paid edition; it is distinct from the free-forever Community Edition.

Platforms

Memgraph supports API access and is available for Linux, macOS and Windows, as well as self-hosted and web use. Cloud instances run on AWS in six regions and range from 1 GB to 32 GB RAM. The platform mix gives teams both local or self-managed deployment paths and a hosted option, though the largest stated Cloud instance is 32 GB RAM.

Who it's for

Memgraph is best for teams building graph applications where streaming inputs, graph algorithms or vector search are central, and for organizations that need enterprise controls such as SSO, role- and label-based access, audit logging and failover. Its named use cases include AI infrastructure, GraphRAG, AI memory, fraud detection, network analysis, supply-chain dependencies and infrastructure monitoring.

It is less compelling for buyers who need a fixed, predictable paid-plan price before sizing their deployment, or who need managed hosting with dedicated engineering support: Cloud includes community support, while dedicated support belongs to Enterprise. Teams unwilling to self-manage should weigh Cloud against that support distinction before choosing.

Pros and cons

Strong free foundation: Community Edition includes ACID transactions, persistence, high-availability replication, streaming connectors, vector search and graph algorithms, so teams can build beyond a bare prototype without a paid plan.

Useful analytics breadth: MAGE’s 60-plus algorithms and integrations with NetworkX, igraph and cuGraph support varied graph-computation workflows.

Good operational depth at Enterprise level: Security controls, SSO, multi-tenancy, failover, disaster recovery, monitoring and direct engineering support address production needs.

Paid costs depend on capacity: Enterprise pricing scales with memory or graph data capacity, so buyers need to size their workload before estimating cost.

Support differs by route: Community Edition and Cloud list community support, while dedicated engineering support is an Enterprise feature.

Alternatives

NebulaGraph is a free, open-source alternative with an Open-source Edition priced at free; consider it when its subset of core features and Apache 2.0 license fit better.

Sparksee is worth considering for mobile deployments, since its listed platforms include Android and iOS; its Academic licence is free for academia with a valid university or school ID.

LadybugDB is another free option, with MIT-licensed source code and pre-compiled binaries; it may suit buyers prioritizing that licensing route.

ArcadeDB offers a free Community plan with its full feature set; consider it if that free-plan scope is a better match.

Aerospike Database is an alternative with a Community Edition described as supporting 8 nodes, 2.5 TB of data and 2 namespaces; consider it when those stated limits match the workload.

OrientDB has a free Community distribution and Basic support priced at €1,000 a month; it may suit buyers who want that explicit support option.

TigerGraph offers a Savanna Free Trial at free, billed at $0 per GB/month* for 24-7 service in US/Tier 1, with up to 32 vCPUs and 256GB RAM per workspace; consider it when those workspace terms fit.

AllegroGraph is another freemium graph platform.

Verdict

Choose Memgraph if your graph workload benefits from streaming connectors, a broad algorithm library and the option to move from a capable free edition to managed or enterprise deployment. The free tier is unusually substantial for teams prepared to run it themselves, while Enterprise adds meaningful security and availability controls. Look elsewhere if your decision depends on a known paid price in advance, or if you need dedicated support without buying Enterprise.

Verdict

Memgraph pairs graph querying and streaming connectors with algorithms, visual tools, and a free Community Edition. Enterprise pricing depends on capacity, while Cloud instance sizes and hosting regions are specified.

Where Specifiction places it

Alternatives to MemgraphAll alternatives →