Getting startedInstallation

Install Memgraph

Memgraph runs in Docker, on Kubernetes, in Memgraph Cloud and as a native Linux package. To try it on your machine first, follow the quickstart, which starts Memgraph with MAGE and Memgraph Lab in Docker.

Installation options

Pick the option that matches where Memgraph will run:

  • Docker or Docker Compose: development, and production on a single machine. Use the memgraph/memgraph-mage image unless you need the slim memgraph/memgraph image without graph algorithms.
  • Memgraph Cloud: a managed instance.
  • Kubernetes: standalone or high-availability deployments with Helm charts.
  • Linux: a native install from the DEB or RPM package, or on Windows through WSL. The package files are also on Direct download links.
  • Source build and Packaging: compile Memgraph yourself or build your own .deb or .rpm package.

Since Memgraph 3.13, there is one .deb package and one .rpm package per architecture. They run on any Linux distribution with glibc 2.31 or newer. Install Memgraph on Linux covers both packages and how to check your glibc version.

Before you run Memgraph, check the system configuration guidelines, especially the vm.max_map_count parameter.

System requirements

Native Linux packages run on any Linux with glibc 2.31 or newer: Debian 12+, Ubuntu 22.04+, RHEL, CentOS and Rocky 9+, and Fedora 42+.

Memgraph requires Python 3.10 or newer on the host, for Python query modules and the auth and SSO modules. Memgraph uses the host’s Python at runtime, so any version from 3.10 up works. On distributions whose default python3 is older (such as CentOS 9 with 3.9), the package installs and uses a newer interpreter. Docker images already include a suitable Python.

The minimum and recommended hardware:

MinimumRecommended
CPUServer or desktop processor:
Intel Xeon
AMD Opteron/Epyc
ARM machines or Apple M1
Amazon Graviton		Server processor:
Intel Xeon
AMD Opteron/Epyc
ARM machines or Apple M1
Amazon Graviton
RAM1 GB≥ 16 GB ECC
Disk1 GBat least 3x the amount of RAM (NVMe SSD, mirroring RAID 1)
Cores1 vCPU≥ 8 vCPUs (≥ 4 physical cores)
Network Throughput100 Mbps≥ 1 Gbps
Network Latency< 10ms< 1ms

The disk is used for storing database durability files, including snapshots and write-ahead logs. By default, Memgraph keeps the three most recent snapshots; set the --storage-snapshot-retention-count flag to change that. A snapshot is usually smaller than the RAM needed to load the same data into memory.

The number of CPU cores you need depends on your workload. More cores let Memgraph run more queries in parallel.

To size RAM for your data, see storage memory usage.

llms.txtDocker