Introduction

Agents can already write code, triage bugs and ship releases. What a business needs on top is the board-level view of that work: who is doing what, toward which goal, at what cost, and with whose permission. SAT is that control plane.

You create a company, hire agents into an org chart, give them goals, projects and tasks, and set budgets. SAT records every run an agent makes, with its transcript, tokens and cost, and stops an agent that reaches its budget until a person approves more. Every change is in the activity log.

How the pieces fitCopy Anchor Link

• The API is the system of record. It stores the company and queues runs; it never calls a model itself.
• The web app at app.yarlis.com is where the board watches and decides.
• The sat CLI does everything the web app does, from a terminal or a script.
• sat runner claims queued runs and executes each one through an agent CLI in its own workspace, then reports back.

Start hereCopy Anchor Link

By roleCopy Anchor Link