>_ kunobi-kobe

Coming Soon

Instant Kubernetes Clusters for CI/CD

Pre-warmed virtual clusters for your pipelines. Claim one in under 5 seconds, use it, release it.

<5sClaim Time

60–70Clusters / 16GB Node

0Secrets Stored

ci-pipeline.sh

>_ how-it-works

Four steps. Zero waiting.

Define once, claim instantly, release cleanly. Kobe handles provisioning, recycling, and pool maintenance in the background.

Define a ClusterPool

Declare the runtime, size, and warm count. Kobe provisions and maintains a ready pool of clusters automatically.

Authenticate via OIDC

Your CI job uses its existing identity — GitHub Actions token, Clerk, or Auth0. No API keys, no secrets to rotate.

Claim & Use

A single API call returns a kubeconfig for a ready cluster. Sub-second bind for pre-warmed pools. Run your workloads immediately.

Release & Recycle

When the job finishes or the TTL expires, Kobe recycles the cluster and refills the pool. No manual cleanup.

Define a ClusterPool

Declare the runtime, size, and warm count. Kobe provisions and maintains a ready pool of clusters automatically.

Authenticate via OIDC

Your CI job uses its existing identity — GitHub Actions token, Clerk, or Auth0. No API keys, no secrets to rotate.

Claim & Use

A single API call returns a kubeconfig for a ready cluster. Sub-second bind for pre-warmed pools. Run your workloads immediately.

Release & Recycle

When the job finishes or the TTL expires, Kobe recycles the cluster and refills the pool. No manual cleanup.

>_ capabilities

Built for scale and security

Policy-as-code authorization, golden snapshots, full observability, and priority queuing — everything you need for production CI/CD infrastructure.

Sub-5s Cluster Claims

Pre-warmed pools mean clusters are already running. Claim binds in under 1 second for ready clusters.

Zero-Secret Auth

OIDC-only authentication. GitHub Actions, Clerk, Auth0 — no API keys stored, no secrets rotation ever.

Policy-as-Code

AccessPolicy CRDs give fine-grained control: per-identity TTL, concurrency limits, and pool access quotas.

Golden Snapshots

Velero integration lets you restore from pre-initialized snapshots instead of cold-provisioning from scratch.

Priority Queuing

FIFO with weighted priorities when pools are exhausted. Requests auto-serve as clusters free up.

Full Observability

Prometheus metrics, OpenTelemetry tracing, and diagnostic bundles to S3 — built in, not bolted on.

>_ runtimes

Choose your runtime

Pluggable backends with no vendor lock-in. From lightweight virtual clusters to full CAPI-managed infrastructure.

vkobe

★ max density

Rust-native virtual clusters. The default runtime for Kobe.

~150–200 MB / cluster

k3s

stable

Lightweight Kubernetes. Battle-tested and broadly compatible.

CNCF-certified

k0s

minimal

Zero-friction Kubernetes. Single binary, minimal dependencies.

single binary

CAPI

enterprise

Cluster API managed. Full infrastructure control end-to-end.

managed lifecycle

Performance & Architecture

Claim latency

Kobe + vkobe: < 5 seconds
vCluster: 10–30 seconds
On-demand provisioning: 30–360 seconds

Density (16 GB node)

Kobe + vkobe: 60–70 clusters
vCluster: 25–30 clusters
On-demand provisioning: 3–5 clusters

Memory per cluster

Kobe + vkobe: ~150–200 MB
vCluster: ~400–500 MB
On-demand provisioning: 2–4 GB

Syncer implementation

Kobe + vkobe: Rust-native
vCluster: Go
On-demand provisioning: N/A

Cloud workload identity

Kobe + vkobe: AWS IRSA · GCP WI · Azure MSI
vCluster: Not supported
On-demand provisioning: Varies

Pre-warming

Kobe + vkobe: Built-in pool management
vCluster: Manual
On-demand provisioning: Not applicable

Stop waiting for clusters

Kunobi Kobe is part of the Kunobi platform. Pre-warm your CI/CD infrastructure and never block a pipeline again.

Learn More Contact Us

Available for:

macOSWindowsLinux