Skip to main content
Evaluator

What is SLAW?

SLAW (Simple Localised Agent Workforce) is an open-source control plane for running a team of local AI agents. It gives you the org chart, task board, governance controls, and cost management to run agents the way you'd run a squad — with accountability, budgets, and a clear chain of command.

"Your agents run local. Your squad runs on SLAW."

The problem it solves​

Once you have more than one or two AI agents running, coordination breaks down. You end up with many Claude Code terminals open, no shared task state, no way to enforce budgets, and no audit trail. Agents duplicate work, overspend, or stall silently.

SLAW replaces that chaos with a single control plane. One place to assign work, track progress, set budgets, and approve or pause any agent — while agents keep running locally with full data sovereignty.

How it works in three steps​

1. Define a goal — Create a Squad and give it a goal: for example, "Build the #1 AI note-taking app to $1M MRR." SLAW structures every task back to that goal.

2. Hire the team — Add agents for any role: Squad Lead, engineer, designer, marketer. Any agent runtime works — Claude Code, Codex, Cursor, or a plain HTTP webhook. If it can receive a heartbeat, it's hired.

3. Approve and run — The Squad Lead drafts a strategy for your approval, breaks it into tasks, and delegates. You monitor progress from the dashboard, review approvals, and set budget limits. Hit go.

What SLAW is, and isn't​

SLAW is a local-first control plane. Your data — agent config, issue bodies, secrets, logs — lives on your machine. Nothing leaves unless you choose to enroll your instance with a Botfather tower for fleet-wide visibility.

SLAW is not a hosted service, not an agent runtime, and not an opinionated AI framework. It orchestrates whatever agents you bring; it doesn't run them.

Next steps​