Skip to main content
Studio runs work you would otherwise do by hand. A flow starts from a trigger, such as a new work item, a call from another system or a time of day, and then runs its steps one after the other. Each step is an action of a plug: create a work item, send a notification, call another system, ask a person to approve, or let an AI agent reach a goal. Studio was called Collabase Automation before.

Why use Studio

Flows live in an Automation workspace that a team owns. The team sees every flow, every run and every change in one list, and a run that fails says which step failed and why.
A flow acts with the rights of its workspace’s service account, never with more. Connection values are stored encrypted and never shown again. Plugs that Collabase did not write run in an isolated sandbox with no route to your internal systems.
A step can ask a named person to approve before the flow goes on. A sensitive Space needs a second admin before a flow is published or switched on. An AI agent can be held to the same rule for each tool it uses.
Everything except the AI steps, building a step from a sentence and the Collabase AI tools works without an AI provider.

Key concepts

How to access Studio

1

Get the Studio right

An administrator gives you the Studio app right under Administration, Users. Anyone with the right can create an Automation workspace.
2

Open Studio

Select Studio in the app drawer. A system administrator finds a default Space already there.
3

Create a flow

Open a Space, select New flow and build it in the editor, or start from a template or a file.

What Studio is made of

They combine freely. An Agent step can start a flow as a tool, a flow can run an agent as a step, and both can read and write the same tables.

Quick start

The first page of Studio builds a flow or an agent from a description and lists everything you may open. See Quick start.

Build with AI

Describe a step in a sentence and let Collabase add it, ask Collabase AI why a run failed, or connect Claude, Cursor or your own agent to the flows of a Space. See Build with AI and the MCP server.

Next steps

Building flows

Triggers, steps and what a run does.

Triggers

Events, webhooks, schedules and chat.

Agent step

Let a model reach a goal with tools.

Template library

Start from a flow that works.

Workspaces

Members, roles, sensitive Spaces and audit.

Build plugs

Connect your own systems.
Last modified on November 4, 2026