Create a task
Add a brand-new task to the project. This only ever adds: it never overwrites existing data, and the new task shows up in the product like any other.
Example: “Create a task titled ‘Write Q3 launch brief’”
Usage: give the new task a title (the only required field). You can also set its description, status, priority, due date, assignee, parent task (to make it a subtask), labels, and any files to attach on creation.
Response: the created task, including its generated id and identifier along with the status, priority, assignee, and other fields it was created with.
Authorizations
Writesonic API key
Headers
Project ID for tenant scoping
Overrides the key's workspace
Body
Task title. Optional (blank/omitted) when a description is given.
500Task body/details.
Initial task status.
backlog, todo, in_progress, pending_human_approval, pending_credits, in_review, done, cancelled Task priority.
urgent, high, medium, low, none Whether the assignee is a human user or an agent.
human, ai_agent, system Required unless assignee_type is system - the user id for human, or the agent's UUID (as a string) for ai_agent.
Parent task, to create this as a subtask.
Content-intelligence page this task relates to, if any.
Market this task is for, if any. Not accepted with parent_task_id: a subtask takes its parent's market.
Action Center item this task was created for, if any. Set by the Action Center only.
Source of the task.
user, agent, content_intelligence Due date for the task.
Labels to attach to the task.
Per-task overrides of project-level agent settings. Value per key: for skip_review, one of ReviewMode (require_review, auto_approve, defer) - not a bool; for gated_tools/gated_capabilities, a dict of tool/capability NAME (not id) -> bool - keys are resolved by name against active tool/capability rows.
Files to attach to the task on creation.
Response
Successful Response
backlog, todo, in_progress, pending_human_approval, pending_credits, in_review, done, cancelled urgent, high, medium, low, none human, ai_agent, system human, ai_agent, system user, agent, content_intelligence