Ross ROSS = Recommend OSS · open-source software intelligence for agents

kcp-dev/kcp

Kubernetes-like control planes for form-factors and use-cases beyond Kubernetes and container workloads. observed · 2026-08-28

github.com/kcp-dev/kcp · homepage · Go · Apache-2.0 (permissive) observed · 2026-08-28

Health v2 · maintenance only

96/100

  • Activity 98
  • Release rhythm 92
  • Longevity 100
How is this computed?

round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-02. Adoption (stars, forks) is never an input.

  • gap_med: 11
  • age_days: 1981
  • days_rel: 58
  • days_push: 12
  • n_releases_24m: 26

Full methodology

Adoption not part of the score

2816 stars · 458 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

kcp is a horizontally scalable, Kubernetes-like control plane that provides many isolated 'workspaces', each acting as its own lightweight Kubernetes-style cluster with its own APIs and RBAC. It is a CNCF Sandbox project written in Go, designed as a building block for massively multi-tenant SaaS platforms offering Kubernetes-native APIs beyond container workloads.

Use cases

  • run many isolated Kubernetes-like clusters cheaply without full cluster overhead
  • build a multi-tenant SaaS platform exposing Kubernetes-native APIs to thousands of tenants
  • offer APIs centrally as a service provider with multi-tenant operators
  • give teams their own workspace with custom CRDs and RBAC
  • host Kubernetes-like APIs for form factors beyond container orchestration
  • shard workspaces across instances for global scale

When to choose

  • you need massively multi-tenant Kubernetes-style APIs without provisioning a real cluster per tenant
  • you are a SaaS or platform team building on Kubernetes APIs at scale
  • you want isolated workspaces as cheap and fast as namespaces
  • you want to complement, not replace, existing Kubernetes clusters

When to avoid

  • you need workload scheduling or running pods, since kcp intentionally has no compute primitives
  • you just need a single standard Kubernetes cluster
  • you need the removed syncer/TMC workload synchronization features
  • you are not comfortable with a CNCF Sandbox-stage project

Facets

service · maturity active

api-framework microservices cloud self-hosted developer-tools cloud-computing infrastructure-as-code microservices self-hosted apis go cross-platform self-hosted kubernetes control-plane multitenancy workspaces saas platform-engineering cncf-sandbox api-management devops linux macos docker

5 sources

Member repositories

RepositoryRoleHealth v2
kcp-dev/kcpmain96

For agents

markdown · JSON · MCP: product_card(name="kcp-dev/kcp")

Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem