Hermes OpenShift: deployment tool for Hermes Agent

Manifests to run Hermes Agent on Red Hat OpenShift AI with vLLM GPU model serving

What Hermes OpenShift does

The repository packages Hermes Agent in a UBI 9-based container image and deploys it beside a KServe vLLM InferenceService on OpenShift AI. Hermes reaches the model over an OpenAI-compatible endpoint. The gateway pod serves HTTP endpoints for health checks and for Telegram, Discord and API traffic, and the README lists Telegram, Discord, Slack, WhatsApp and Signal as supported messaging platforms.

Numbered manifests create the hermes namespace, a 50Gi volume for vLLM model weights, a KServe vLLM runtime with CUDA support, the InferenceService, a 10Gi volume for Hermes data, a ConfigMap holding the vLLM endpoint, the Hermes deployment with health probes, a ClusterIP service on port 8080 and an optional TLS Route. The setup complies with the restricted-v2 Security Context Constraint, and one oc apply -k manifests/ deploys both the model serving and the agent.

Key features

UBI 9-based container image with the Hermes Agent stack
Direct integration with a vLLM InferenceService on OpenShift AI
Messaging gateway for Telegram, Discord, Slack, WhatsApp and Signal
Kubernetes health probes and lifecycle management
Compliance with the restricted-v2 Security Context Constraint
Persistent volumes for Hermes skills, memories and user models

When to use it

Running Hermes Agent on an enterprise OpenShift AI cluster
Serving a self-hosted Qwen model with vLLM for the agent
Exposing the Hermes gateway to chat platforms through an OpenShift Route

Who it is for: Platform engineers who run Red Hat OpenShift AI and want Hermes Agent deployed against their own GPU model serving.

How it fits with Hermes Agent

It is a deployment template built specifically around Hermes Agent, its gateway and its persistent storage.

How to install Hermes OpenShift

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

git clone https://github.com/aicatalyst-team/hermes-openshift
cd hermes-openshift
oc apply -k manifests/
oc get pods -n hermes -w

Requirements: Red Hat OpenShift 4.x cluster with OpenShift AI, the oc CLI authenticated to the cluster, 10Gi of storage, and a GPU node for vLLM (optional but recommended)

Note: GitHub reports the repository's license as NOASSERTION, so check the license terms before reusing the manifests.

FAQ

What is Hermes OpenShift?

Hermes OpenShift is a set of manifests and a container setup for deploying Hermes Agent on Red Hat OpenShift AI. It connects the agent to a GPU-accelerated vLLM model server.

Does Hermes OpenShift work with Hermes Agent?

Yes. It is built specifically to run Hermes Agent, including its messaging gateway, learning loop, skills storage and cron scheduler.

What do I need to run Hermes OpenShift?

You need an OpenShift 4.x cluster with OpenShift AI, an authenticated oc CLI and 10Gi of available storage. A GPU node for vLLM is optional but recommended.