The prefect-client package is a minimal-installation of prefect which is designed for interacting with Prefect Cloud or remote any prefect server. It sheds some functionality and dependencies in exchange for a smaller installation size, making it ideal for use in lightweight or ephemeral environments. These characteristics make it ideal for use in lambdas or other resource-constrained environments.

prefect-client shares the same installation requirements as prefect. To install, make sure you are on Python 3.9 or later and run the following command:

pip install prefect-client

Next, ensure that your prefect-client has access to a remote prefect server by exporting the PREFECT_API_KEY (if using Prefect Cloud) and PREFECT_API_URL environment variables. Once those are set, use the package in your code as you would normally use prefect!

For example, to remotely trigger a run a deployment:

from prefect.deployments import run_deployment


def my_lambda(event):
    ...
    run_deployment(
        name="my-flow/my-deployment",
        parameters={"foo": "bar"},
        timeout=0,
    )

my_lambda({})

To emit events in an event driven system:

from prefect.events import emit_event


def something_happened():
    emit_event("my-event", resource={"prefect.resource.id": "foo.bar"})

something_happened()

Or just interact with a prefect API:

from prefect.client.orchestration import get_client


async def query_api():
    async with get_client() as client:
        limits = await client.read_concurrency_limits(limit=10, offset=0)
        print(limits)


query_api()

By design, prefect-client omits all CLI and server components. This means that the CLI is not available for use and attempts to access server objects will fail. Furthermore, some classes, methods, and objects may be available for import in prefect-client but may not be "runnable" if they tap into server-oriented functionality. If you encounter such a limitation, feel free to open an issue describing the functionality you are interested in using and we will do our best to make it available.

There's lots more you can do to orchestrate and observe your workflows with Prefect! Start with our friendly tutorial or explore the core concepts of Prefect workflows.

Prefect is made possible by the fastest growing community of thousands of friendly data engineers. Join us in building a new kind of workflow system. The Prefect Slack community is a fantastic place to learn more about Prefect, ask questions, or get help with workflow design. All community forums, including code contributions, issue discussions, and slack messages are subject to our Code of Conduct.

See our documentation on contributing to Prefect.

Thanks for being part of the mission to build a new kind of workflow system and, of course, happy engineering!