The Pipekit CLI lets you interact with Pipekit services without leaving your terminal.
The pipekit CLI is the command-line surface for Pipekit. Everything the UI does is also doable from the CLI: submit Workflows, list Runs, stop or restart a Run, manage CronWorkflows, import from an Argo Workflow Archive.
The CLI also hosts the MCP Server that lets LLM clients (Claude Desktop, Cursor, Claude Code, Goose) call Pipekit on your behalf.
Install: how to install the CLI on macOS, Linux, Windows, and NixOS.
Commands: the core command surface (login, submit, get run, list, stop / terminate / restart, logs, update cluster, import workflows).
Using with Hera: generate tokens for Hera workflows and pair the CLI with the Pipekit Python SDK.
Advanced: Docker container usage, non-interactive login, token storage, log-level flags, SBOM extraction.
Cron Workflows: CronWorkflow lifecycle commands.
MCP Server: pointer to the AI section.
Release notes are published on the Pipekit Releases site.
# Install (macOS / Linux)
brew install pipekit/tap/cli
# Log in
pipekit login
# List clusters you have access to
pipekit list clusters
# Submit a workflow
pipekit submit --cluster-name=<cluster> path/to/workflow.yaml
# View logs for a Run
pipekit logs --run-uuid=<uuid> --followLast updated