AiTask

Type: Object

Represents a single, executable step within an AI-driven workflow. When an AI 'Execution' (a running process) begins, it generates a series of these tasks from a predefined template. Each task is a concrete action to be performed, such as 'Analyze customer artwork for preflight errors,' 'Calculate ink consumption for job #12345,' or 'Dispatch shipping notification.' This resource tracks the lifecycle of that action—from pending to in-progress, completed, or failed. It's the fundamental unit of work in the AI system, providing granular visibility and control over the automated processes that drive the print facility.

Fields

Name	Type	Args	Description	Deprecated
agent	AiAgent!
agentId	ID!
completedAt	DateTime		When the task completed (successfully or with failure)
context	String		Additional context for the task
createdAt	DateTime!		When the task was created
dependencies	[AiTask!]!	filter: AiTaskFilterInput limit: Int offset: Int sort: [AiTaskSortInput]
description	String!		Interpolated description with actual values from the execution input
errorMessage	String		Error message if the task failed
execution	AiExecution!
executionId	ID!
expectedOutput	String		Description of what output is expected from this task
id	ID!
metadata	JsonString		Additional metadata for the task
outputJsonSchema	JsonString		JSON Schema for structured output. When provided, the LLM will use JSON mode with this schema to ensure valid JSON output
priority	Int		Task priority (higher numbers = higher priority)
result	String		The output produced by the agent
startedAt	DateTime		When the task started execution
state	String!		The current state of the task: pending, assigned, in_progress, completed, failed, or cancelled
taskTemplate	AiTaskTemplate		The template this task was created from
taskTemplateId	ID
updatedAt	DateTime!		When the task was last updated

Relationships

• AiAgent
• AiExecution
• AiTaskTemplate