Working with Toone

Approvals, permissions and audit trail for AI agents

How Toone keeps people in control: approval gates inside each routine, scoped permissions per agent, a receipt for every action, and local-first data handling.

September 28, 20262 min readRoutine design guide

AI agent routines: recurring work written in plain English

What an AI routine is, how to write one in natural language, schedule it, add approval steps and review each run in Toone. Not a habit tracker.

September 28, 202615 min readWorking with Toone

Make recurring work easier

Begin with a useful workflow and a result your team can review.

September 28, 20262 min readOrganization selection guide

How to Choose an AI Agent Organization Template

Choose an AI agent organization pattern by job, roles, permissions, shared data, handoffs, evidence, recovery, maintenance, and non-fit.

September 28, 202615 min readExplore Toone

Start with one workflow

Start with one recurring workflow: pick the work worth repeating, give the agent the context it needs, and review the first runs before you widen the scope.

September 28, 20262 min readExplore Toone

Work with specialised agents

How specialised agents divide recurring work in Toone: roles with scoped permissions, routines with fixed steps, handoffs between agents, and a reviewed result.

September 28, 20262 min readExplore Toone

Keep useful context close

Keep the context your agents need close to the work: what to write down, where it lives on your Mac, and how routines pick it up on every run.

September 28, 20262 min read