Great question! In Prefect 3.x, you set default flow parameters at deploy time with the

prefect deploy

CLI using either

--param

(repeatable key=value) or

--params

(single JSON object). These are mutually exclusive. - Use

--param key=value

for individual parameters (repeatable). Values are parsed as JSON if possible; otherwise treated as strings. - Use

--params '{"k":"v"}'

to provide all parameters at once as a JSON object. Examples:

# Set individual params (repeatable)
prefect deploy ./flows.py:my_flow \
  --name my-deployment \
  --param customer_id=12345 \
  --param enabled=true \
  --param tags='["gold","priority"]'

# Set all params at once (JSON)
prefect deploy ./flows.py:my_flow \
  --name my-deployment \
  --params='{"customer_id":12345,"enabled":true,"tags":["gold","priority"]}'

Notes: - Don’t mix

--param

and

--params

in the same command (you’ll get an error). - These values become the deployment’s default parameters. You can override them at run time:

prefect deployment run my-flow/my-deployment --param customer_id=999

If you prefer YAML, you can put defaults under

parameters:

for a deployment in `prefect.yaml`:

deployments:
  - name: my-deployment
    entrypoint: flows.py:my_flow
    parameters:
      customer_id: 12345
      enabled: true

CLI flags will merge with and override values defined in

prefect.yaml

. Programmatic equivalent (Python):

from prefect import flow

@flow
def my_flow(customer_id: int, enabled: bool = True): ...

flow.from_source(
    source="<https://github.com/org/repo.git>",
    entrypoint="flows.py:my_flow"
).deploy(
    name="my-deployment",
    work_pool_name="my-pool",
    parameters={"customer_id": 12345, "enabled": True}
)

Docs: - Create deployments - prefect.yaml - Trigger deployment runs