Run is where a Playbook earns its keep: YAMO executes it with your tools and your data, on demand or on a schedule, unattended, and shows you exactly what it did.

Inputs

A Playbook declares its inputs in its own document, so the run form is never a surprise:

Run inputs are asked on every run. The URL SEO Audit template asks for one thing: the page to audit.
Config inputs are set once by you and reused: part of setup, not of each run.
An input can carry a default. Scheduled runs use the defaults when nobody is there to type: a recurring Playbook ships its defaults precisely so its schedule can run without you. Set them to your own values before switching a schedule on.

Starting a run

Deploy the Playbook to turn its Draft Service into a private one you can run. Until you deploy, nobody can run it, including you. Open the Service, create or reopen an Instance, then choose Run Service. The Instance conversation asks for each required run input, shows the exact price if one applies, and asks you to confirm before it starts. Runs execute in the background: closing the tab changes nothing. The deliverable lands in the Instance's work rail.

Schedules

Recurring Playbooks carry a cadence in their document, such as "Weekly, Monday 07:00". Pause and resume that automation from the Playbook's settings (Schedule tab, behind the header cog). Each scheduled execution is still a normal Run with the same stages, run folder, and receipt; manual Runs start from a Service Instance.

Notifications

You don't have to watch. Email notifications cover the three moments that matter when you're away:

Awaiting approval: a Play with a human gate has paused the run. The email links back to the Playbook to review; nothing is approved from the email itself.
Succeeded: the run finished, with a link to the output.
Failed: the run stopped, with the honest reason.

The run folder

Every run keeps its evidence. From any output you can open the run folder and walk back through:

the inputs snapshot: exactly what this run was given,
each stage's output in sequence: the work between the input and the deliverable,
the receipt, itemised with the LLM cost and the operation charges as separate lines. Token usage is shown as a transparency readout: you are never billed "per token", and the LLM line is passthrough at cost. Run costs are rounded up to the nearest penny.