Great question. In Prefect 3.x you can determine this at runtime by inspecting the current flow run’s context. The two most useful fields are: - `flow_run.auto_scheduled`: True when the scheduler created the run from a deployment’s schedule - `flow_run.created_by.type`: Typically 'scheduler' (scheduled), 'user' (manual via UI/CLI), 'automation', or 'api' Quick example inside your flow:

from prefect import flow
from prefect.context import get_run_context

@flow
def my_flow():
    fr = get_run_context().flow_run

    if fr.auto_scheduled:
        trigger = "schedule"
    else:
        trigger = "manual or other (e.g. automation/API)"

    who = getattr(fr.created_by, "type", None)  # 'scheduler', 'user', 'automation', 'api', etc.

    print(f"Triggered by: {trigger} (created_by.type={who}, deployment_id={fr.deployment_id})")

Notes: - If you click “Run” on a deployment in the UI (or use CLI to run a deployment),

deployment_id

will be set but

auto_scheduled

will be False and

created_by.type

will generally be 'user' → that’s a manual run of a deployment. - If the scheduler created the run from the deployment’s schedule,

auto_scheduled

will be True and

created_by.type

will typically be 'scheduler'. - Ad-hoc runs (not from a deployment) will have

deployment_id=None

. If you just want a simple boolean for “scheduled vs not”:

from prefect.context import get_run_context

def is_scheduled_run() -> bool:
    return get_run_context().flow_run.auto_scheduled

Optional runtime helpers: -

prefect.runtime.deployment.id()

returns the current deployment id (None for ad-hoc runs) -

prefect.runtime.flow_run.parent_flow_run_id()

is set for subflows -

prefect.runtime.flow_run.parent_deployment_id()

can help when nesting is involved Docs: - Access runtime info - Runtime context - prefect.runtime.flow_run API - FlowRunResponse schema Do you also need to distinguish “manual” vs “automation rule” vs “API call”? If so, we can branch on

created_by.type

more explicitly.