Skip to navigation

Factory overview

A factory groups the work your team wants to run. Lines define a trigger, stages, and the route between stages. Agents do the work. Each Agent has a Machine, the prepared environment it uses for its sessions. A Machine can serve Agents on several Lines.

The work model

ConceptWhat it controls
factoryThe scope for Lines, Agents, Machines, and related resources
LineThe trigger, stages, transitions, and limits for repeatable work
AgentA reusable prompt or script worker configuration assigned to a Machine
MachineReusable repository, tools, and runtime configuration for an Agent
IntegrationsConnections to tools such as GitHub, Slack, and Linear for incoming work and Agent actions
GatewaysAccess policies and credential injection for Agents calling connected services
Manager and DelegatorRouting decisions and delegation while a Line runs

A Line run records one execution. Inspect its stages and events to see what each Agent did and where work stopped. See Lines for the manifest and factory troubleshooting for failures.

Start and inspect work

Choose a manual, scheduled, webhook, or integration trigger. Ask a local agent using the Islo skill to prepare and deploy the Line. The manual path is useful while you test it:

islo factory line run <line-name> --param input="Review pull request 123"
islo factory line-run status <run-id>
islo factory line-run events <run-id>

To start with a complete example, follow deploy your first Line. To prepare the Agent’s environment first, see Machines.

Platform capabilities

Machine execution uses isolated sandboxes, with setup scripts and snapshots for preparing and restoring environments. These are runtime infrastructure details within Factory workflows. The advanced references also cover direct management through the CLI and API.