Yes. In Prefect 3.x, calling one flow from inside another (e.g.,

ret = B()

inside

A

) creates a subflow run. What that means: - Execution: Runs in-process with the parent flow, synchronously.

A

will wait for

B()

to finish and

ret

is the return value of

B

. - UI/lineage: You’ll see a nested “subflow” under the parent flow run in the Prefect UI (the child run has the parent’s run ID). - Errors: Unhandled exceptions in

B

propagate to

A

and can fail the parent run. - Parameters: Pass arguments to

B

like a normal Python function; Prefect records them for the subflow run. If you want

B

to run as an independent flow run (e.g., on different infrastructure or not nested in the UI), deploy

B

and trigger it via

run_deployment

. By default,

run_deployment(..., as_subflow=True)

will still nest under the caller; set

as_subflow=False

to decouple. Example:

from prefect import flow
from prefect.deployments.flow_runs import run_deployment

@flow
def B(x: int = 1) -> int:
    return x + 1

@flow
def A():
    # Subflow (in-process, nested in UI)
    r1 = B(2)

    # Separate deployment run (still nested by default)
    child = run_deployment(
        name="B/my-deployment",
        parameters={"x": 3},
        as_subflow=True,  # set to False to make it a top-level run
    )
    return r1

Docs: - Running deployments (includes

as_subflow

): Run deployments - Creating deployments: Create deployments