Hyperspool docs

Hyperspool is a durable workflow engine with a CI layer on top. Runs sleep through soak windows, wait for approvals and external signals, and spawn child runs - and the same substrate checks out your repo, runs your build, and reports your commit status. It resumes exactly where it left off across agent restarts, server restarts, and weeks of waiting.

Where to start

• Quickstart - install the spool binary and get a first run green in two commands, entirely on your machine.
• Concepts - the mental model, one page.

Then whichever of the three you’re writing in - they run on one engine and compose with each other, so this is a matter of what fits the work, not a ladder:

• YAML pipelines - the pipeline dialect: jobs, steps, durable waits, approval gates, child workflows.
• Polyglot scripts - the same workflows in bash, Python, TypeScript, or anything that can spawn a process.
• Authoring on the canvas - build a graph of nodes and edges in the browser, with live validation and expression autocomplete.

Hosted, in three steps

Point the CLI at the server, start an agent on a machine that will run your jobs, then submit. The middle one is the step to not skip: jobs run on machines you provide, so with no agent running, a submitted run waits for a machine that never arrives.

The commands are in Getting started, and Deploying agents covers systemd, macOS, containers and self-update.

From there spool runs, spool logs and everything in the CLI reference work against the server.