Skip to main content

FAQ

Quick answers to the questions that come up most — before you adopt SLAW and once you're running it.

If a term here is new, the Concepts glossary is the canonical definition home. This page links to deeper material rather than repeating it.

Getting started​

What is SLAW, in one sentence?​

A self-hosted control plane that orchestrates a squad of AI agents — org chart, goals, budgets, and governance — running locally, with a central control tower over the whole fleet. See What is SLAW.

What does a typical setup look like?​

Locally, a single Node.js process manages an embedded PostgreSQL database and local file storage — no separate database to install. For production you point SLAW at your own Postgres and deploy however you like. You configure projects, agents, and goals; the agents take the work from there. The fastest path is one command:

npx @slaw-ai/slaw onboard --yes

This starts the control plane on http://localhost:3100. See the Quickstart.

Do I need a SLAW account or cloud service?​

No. SLAW is open source and self-hosted, and the quickstart needs no account. Everything runs on your machine by default.

Running a squad​

Can I run more than one squad?​

Yes. A single deployment can run an unlimited number of squads with complete data isolation between them.

Do agents run continuously?​

By default, agents wake on scheduled Heartbeats and on event triggers such as task assignment and @-mentions. You can also wire in continuously running agents — you bring the agent, and SLAW coordinates it.

Which agents and providers work with SLAW?​

SLAW is provider-agnostic: Claude Code, Codex, Cursor, Gemini, bash and HTTP/webhook bots, and custom Adapters. The rule of thumb from the project: if it can receive a heartbeat, it's hired.

How do I stop an agent from overspending?​

Set a monthly budget per agent. When an agent reaches its limit it is auto-paused on its next heartbeat and queued work is cancelled, so a runaway loop stays bounded rather than open-ended. See Costs & budgets.

How SLAW compares​

How is SLAW different from an agent like Claude Code?​

SLAW uses those agents. It orchestrates them into a squad with an org chart, budgets, goals, governance, and an audit trail — the coordination layer above any single agent.

Why not just point my agent at Asana or Trello?​

Agent orchestration has subtleties a generic task board doesn't handle: who has work checked out, maintaining sessions across runs, tracking cost, and enforcing governance. SLAW does that coordination for you. (Bring-your-own ticket system is on the roadmap.)

What is SLAW not?​

It is not a chatbot, an agent framework, a workflow builder, or a code-review tool. It models a squad of agents — see SLAW vs. alternatives for the full comparison.

Data and privacy​

What leaves my machine?​

Agent work, issue content, prompts, and secrets stay local by default. For the full picture — local-first storage, key handling, and what an optional fleet control tower can see — read Security & sovereignty.

Next steps​