Run AI on your infrastructure, and know what it did
Olympusx AI serves models, grounds agents on your knowledge and gives every team a gateway, on clusters you own. These pages cover installing it, running it, and every view in the console.
Start here
Install and run the platform
Use the workspace
What Olympusx AI isThe platform and the workspace, who uses each, and what stays on your network.
How it fits togetherThe console, the gateway, tenants and clusters, and what crosses between them.
QuickstartFrom a fresh install to a first tenant, a first model and a first answer.
Finding your wayThe console's rail, what each role sees, and the command palette.
KubernetesInstall on a conformant Kubernetes cluster with the operator or the Helm chart.
OpenShiftInstall from the operator catalog, with routes and restricted security contexts.
Oracle OKEInstall on Oracle Kubernetes Engine, CPU or GPU node pools.
Air-gappedMirror the images and models, and install with no route to the internet.
Platform administrationClusters, model serving, tenants, billing, identity, backups and retention.
UpgradeMove to a new release, and what to check after.
Code AssistantYour developers keep their editor; models, keys and code stay inside.
SecurityCode assessments and log investigations that cite their evidence and wait for a person.
BuildModels, knowledge, agents, tools, skills, assistants and the playground.
Operate and evaluateRuns, approvals, conversations, traces, the review queue and scores.
GovernGuardrails, users and roles, keys and budgets, usage, audit and erasure.
APIThe OpenAI-compatible gateway: authentication, endpoints, errors and limits.