For the complete documentation index, see llms.txt. This page is also available as Markdown.

Pipekit Cloud vs Self-Hosted

Pipekit Cloud and Self-Hosted Pipekit deployment models, side by side.

Pipekit ships in two deployment shapes. Both run the Pipekit Agent inside your Kubernetes Cluster; they differ in where the control plane (UI, API, database) lives.

Component
Pipekit Cloud
Self-Hosted

Control plane

Hosted by Pipekit

Hosted by you (Kubernetes via Helm)

Pipekit Agent

Runs in your cluster

Runs in your cluster

Argo Workflows

Runs in your cluster

Runs in your cluster

Workflow data

Stored in your cluster; metadata in Pipekit Cloud

Stored entirely in your infrastructure

Logs

Optional: Pipekit Cloud or your own backend

Your own backend (optional Pipekit log storage)

Identity provider

Configured per Organization

Configured per Organization

Network connection

Outbound only (agent → Pipekit Cloud via Redis)

All in-cluster; can run fully air-gapped

Choose Pipekit Cloud for the fastest evaluation path and lowest operational overhead. Choose Self-Hosted when your security, compliance, or air-gapping requirements rule out a hosted control plane.

To evaluate Pipekit Cloud, see Get Started > Evaluate Pipekit Cloud. To deploy Self-Hosted, see Self-Hosting Pipekit.

Last updated