Hume — Helm Chart by GraphAware

Hume is an enterprise graph analytics platform that lets teams build, explore, and operationalize graph-based intelligence on top of Neo4j.

This repository contains the official Helm charts for deploying Hume on Kubernetes.

---

Prerequisites

• Kubernetes 1.19+
• Helm 3.2.0+
• A GraphAware account with access to docker.graphaware.com
• A Hume licence key

Neo4j is not included in the chart. You must provision it separately.

Chart version: the commands below install 3.2.0, the latest released version at the time of writing. This branch's hume-helm/Chart.yaml is currently at 3.3.0-dev — an unreleased, in-progress build that has not been published to docker.graphaware.com and cannot be installed. Always use the latest released version (no -dev/-SNAPSHOT suffix); check with your GraphAware contact if unsure.

---

Quick Install

This walks through a local/evaluation setup with port-forwarding — no ingress or domain required. For a full walkthrough with pod-readiness checks and verification steps, see the Quick Start guide.

1. Log in to the Helm registry

helm registry login -u '<username>' -p '<password>' docker.graphaware.com

2. Create a namespace and image pull secret

kubectl create namespace hume

kubectl create secret docker-registry graphaware-docker-creds \
  --docker-server='docker.graphaware.com' \
  --docker-username='<username>' \
  --docker-password='<password>' \
  -n hume

3. (Optional) Pre-load your licence key

kubectl create secret generic hume-licence \
  --from-literal=hume.licence.key=<content-of-your-b64-file> \
  -n hume

If omitted, Hume will prompt you to upload the licence on first login.

4. Install

helm install hume oci://docker.graphaware.com/public/hume --version 3.2.0 -n hume

5. Wait for pods to be ready

kubectl get pods -n hume -w

All pods should reach Running within a few minutes (the API pod may restart once while waiting for the database — this is expected).

6. Access the UI (local / no ingress)

kubectl port-forward service/hume-web 8081:8081 -n hume &
kubectl port-forward service/hume-api 8080:8080 -n hume &

Open http://localhost:8081 and log in with admin@hume.k8s / password.

For evaluation only. This setup uses default admin credentials, embedded PostgreSQL with no replication, and no TLS. It is not suitable for production. Change the default admin credentials before exposing Hume externally, and see Basic Installation for a production-ready setup.

---

Documentation

The full documentation lives in docs/.

I want to…	Go to
Understand what gets deployed	Architecture Overview
Run Hume locally in minutes	Quick Start
Deploy to a real cluster with ingress	Prerequisites → Basic Installation
Configure AWS ALB or Nginx ingress	Ingress
Connect to an external PostgreSQL	Databases
Set up SSO with Keycloak	Authentication
Enable LLM integration	Maestro
Enable alerting	Alerting
Upgrade from a previous version	Upgrade Guide
Set up monitoring with Prometheus	Monitoring
Diagnose a broken deployment	Troubleshooting
See all configuration parameters	Values Reference

---

Charts

Chart	Description
hume-helm	Core platform: Web, API, Orchestra, EventStore, Media, Maestro
hume-alerting	Alerting engine: Controller, Operator, Kafka (separately licensed)

# Install a specific version
helm install hume oci://docker.graphaware.com/public/hume --version 3.2.0 -n hume -f values.yaml

# Upgrade
helm upgrade hume oci://docker.graphaware.com/public/hume --version 3.2.0 -n hume -f values.yaml

# Uninstall (PVCs are retained)
helm uninstall hume -n hume

---

Deployment Scenarios

The deployment-scenarios/ directory contains ready-to-use values files for common setups:

Scenario	Description
external-persistence/	External PostgreSQL and Keycloak
hume-sso-keycloak/	SSO with internal Keycloak
orchestra-cluster/	Orchestra in cluster mode with Prometheus monitoring
maestro/	Maestro with OpenAI
initial-api-key/	Pre-configured API key for automation

---

License

Copyright © 2022 GraphAware

Licensed under the Apache License, Version 2.0. See LICENSE for details.