Skip to main content

WaveMaker Enterprise AI Prerequisites

You can set up WaveMaker Enterprise AI on any machine.

note

This document uses words like VM, Instance to refer a machine.

WME AI setup system requirements

WaveMaker Enterprise AI can be installed on any machine that meets the following requirements. Before you start setting up WaveMaker Enterprise AI, review the minimum and recommended system requirements for each instance type.

WME AI Platform Instance

RequirementMinimum configuration
Memory
  • Minimum 32 GB
CPU
  • 8-core, single CPU system
Hard disk
  • Minimum 450 GB to be allocated
  • For volume-based setups, allocate:
    • 100 GB for /
    • 200 GB for /wm-data
    • 150 GB for /wm-runtime
Host OS
  • Ubuntu 22.x LTS or RHEL 8.x/9.x
  • Kernel 4.4 or later
  • x86 architecture
Software
  • Docker 28.x
  • Python 3.5 or later
  • wget
  • jq
Network

WME AI StudioWorkspace Instance and AppDeployment Instance

RequirementMinimum configuration
Memory
  • Minimum 32 GB
CPU
  • 8-core, single CPU system
Hard disk
  • Minimum 300 GB to be allocated
  • For volume-based setups, allocate:
    • 100 GB for /
    • 200 GB for /data
Host OS
  • Ubuntu 22.x LTS or RHEL 8.x/9.x
  • Kernel 4.4 or later
  • x86 architecture
Software
  • Docker 28.x
  • Python 3.5 or later
  • wget
  • jq
Network

WME AI Observability Instance

RequirementMinimum configuration
Memory
  • Minimum 16 GB
CPU
  • 4-core, single CPU system
Hard disk
  • Minimum 200 GB to be allocated
  • For volume-based setups, allocate:
    • 200 GB for /
Host OS
  • Ubuntu 22.x LTS or RHEL 8.x/9.x
  • Kernel 4.4 or later
  • x86 architecture
Software
  • Docker 28.x
  • Docker Compose 28.x
  • Python 3.5 or later
  • wget
  • jq
Network

IP Addressing and DNS Mapping

You will be needing IP Addresses for the following.

IP Address

  • One static IP for accessing the platform machine from your developer's network.
  • Machine Static IP: This is the IP assigned to the machine during setup and should be accessible on your network, or
    • In the case of VM, it will be the local IP address, which should be rout table from in your LAN.
    • In case of AWS instance: Private static IP for the instance within your VPC (assigned via eth0 or via ENI on eth1,ens5)

DNS Mapping

A DNS domain is mandatory for the Platform instance — developers access WaveMaker Studio using a domain name, not an IP address. DNS for other instances is optional but recommended.

DomainDomain URLRequiredDescription
WaveMaker Studiowavemakerai.[mycompany].comMandatoryUsed to access WaveMaker AI Studio
WaveMaker Deployed Appswmai-apps.[mycompany].com   wmai-stage.[mycompany].com   wmai-live.[mycompany].comOptionalUsed to access WaveMaker AI Studio apps deployed onto WaveMaker AI Cloud
WaveMaker AI Observabilitywmai-analytics.[mycompany].comOptionalUsed to access WaveMaker AI Analytics service. If not configured, use port 5050 via IP.
note

In the preceding table, [mycompany] is used as an example. Replace [mycompany] with your actual domain name.

Docker Container Access

  • An IP range to be assigned to the Docker containers internally. The Minimum CIDR (Classless Inter-Domain Routing) range for Docker container network is 24.

You will be needing to assign a /24 CIDR to Docker during setup. This IP range should not be in use anywhere on your network and can be completely different from your network's range. These IPs are assigned internally by Docker to containers and these IPs won't be exposed on your network.

For example, if your network is using a 10.x.x.x_range and the range_192.168.x.x is not used anywhere in your network, you may assign this 192.168.x.x range to Docker. See here for the possible LAN IP ranges.

Port Requirements

External Access Ports

These ports must be accessible from outside the WME platform network — from developer machines, DevOps teams, and admin machines.

note

Ports 443 on the Platform, AppDeployment, and Observability instances are accessed through a DNS name or load balancer, not directly via IP:port. Ensure the DNS entries (see DNS Mapping) resolve to the respective instances and that traffic on port 443 can reach them through your network or load balancer.

PortInstanceDNS NameAccessed ByPurpose
443Platformwavemakerai.[mycompany].comDeveloper machinesHTTPS access to WaveMaker Studio
443AppDeploymentwmai-apps.[mycompany].com
wmai-stage.[mycompany].com
wmai-live.[mycompany].com
Developers / end usersAccess to deployed WaveMaker applications
443Observabilitywmai-analytics.[mycompany].comDevOps machinesAI observability UI — traces and analytics
5050ObservabilityIP-based, no DNS requiredDevOps machinesFallback access when DNS is not configured for the Observability instance
8080PlatformIP-basedAdmin machinesWaveMaker config portal
22All instancesIP-basedAdmin machinesSSH access for installation and management

Internal Communication Ports

All communication listed here is between WME instances within the platform's private network. None of these ports need to be accessible from outside the WME network.

Recommended: Allow unrestricted communication between all WME instances within the platform's private network.

If your security policy requires restricting traffic to specific ports, open only the ports listed in the following tables.

Open on the Platform Instance — for access from StudioWorkspace and AppDeployment instances:

PortPurpose
443HTTPS access to the Platform Instance
5000Platform services
8500Service discovery
22SSH access
8081Platform communication
2200Container SSH access
8100StudioWorkspace and AppDeployment communication
9200Search and observability services
8000-8020Platform-managed application services
8094AI service communication
8079AI service communication
5432Database connectivity
5433Vector database access for AI features
8083AI Studio and agent-server LiteLLM proxy communication
8086AI Studio and agent-server key management

Open on StudioWorkspace and AppDeployment instances — for access from the Platform Instance:

PortPurpose
22SSH access
2375Docker API access
80HTTP access
5000Platform service communication
8100StudioWorkspace and AppDeployment communication
8888Workspace service communication
9101, 9102, 9100Metrics collection
9404Metrics export
2200-2299Container SSH access
8001-8099Platform-managed application services
3300-3399Database and service communication
9500-9599Platform-managed service communication
3000Routing traffic to AI Studio
3001Routing traffic to AI Studio NGINX
3002Routing traffic to agent-server
5010Backend MCP
5020UI MCP

Open on the Observability Instance — for access from the Platform and all StudioWorkspace instances:

PortPurpose
3000Langfuse — AI trace data forwarding from the Platform and StudioWorkspace instances to the Observability instance

Network Communication

WME instances communicate in two ways:

  • External access — Developer machines access WaveMaker Studio and deployed applications via port 443 using DNS names. DevOps teams access the Observability UI via port 443 (DNS) or port 5050 (IP fallback). Admin machines connect to all instances over port 22 (SSH) and to the Platform over port 8080 (config portal). See External Access Ports.
  • Internal communication — All WME instances communicate with each other within the platform's private network over the ports listed in Internal Communication Ports. None of these are exposed externally.

The following diagram shows the network communication between all WME instances and external access points.

Capacity Planning

WME AI capacity scales horizontally — add more StudioWorkspace or AppDeployment instances to support more concurrent developers or deployments.

Studio Workspace — each 32 GB StudioWorkspace Instance supports the following number of concurrent developer logins, depending on app type:

Application TypeConcurrent developer logins per instance
WEB18
App-Preview-ESBuild18
App-Preview-expo4

AppDeployment — each 32 GB AppDeployment Instance supports up to 20 concurrent app deployments.

note

Capacity is also governed by your license terms — the number of apps that can be developed or deployed cannot exceed what your license allows, regardless of infrastructure size. Add separate instances for each stage in your release pipeline.

WME AI Setup Artifacts

WaveMaker provides the installation artifacts — installer files and images — required to set up WME AI. Before running the installer, ensure each machine is prepared with the OS, Docker, and other software listed in the system requirements above.