Welcome to the docs

Find the right prokube documentation for interactive development, agent workflows, ML workloads, and platform operations.

prokube is a Kubernetes-native AI platform for teams that need to run AI workloads on infrastructure they control. It builds on established open-source tools and integrates them into one platform with a common login, shared workspaces, storage, and observability. Administrators get tested upgrade paths for the whole stack.

Start with LabsInstallationGet enterprise support

Using a prokube release before 1.9?

See the legacy documentation. It covers the platform before 1.9 and does not reflect the behavior described on this site.

Browse by Area

Interactive developmentLabsLaunch JupyterLab, VS Code, RStudio, OpenCode, or custom images.Agents and toolsAgentOpsBuild, evaluate, run, and manage agents on Kubernetes.Model lifecycleMLOpsRun pipelines, track experiments with MLflow, and serve models with KServe.Shared foundationPlatformUnderstand workspaces, API access, observability, and storage.

Common Tasks

Launch a LabStart JupyterLab, VS Code, RStudio, OpenCode, or a custom notebook image inside your workspace.Learn more ->Work with file storageUse S3-backed file storage and PVC-backed volumes for datasets, artifacts, Lab files, and workload data.Learn more ->Configure API accessCreate scoped API keys for SDKs, automation, and external access to platform APIs.Learn more ->Run a pipelineWrite, compile, and submit Kubeflow Pipelines from a Lab to automate reproducible ML workflows.Learn more ->Deploy a modelServe trained models as inference endpoints with KServe, test them in the UI, and access them via API.Learn more ->