Introduction
What is Prefect?
Prefect is an open-source orchestration engine that turns your Python functions into production-grade data pipelines with minimal friction. You can build and schedule workflows in pure Python—no DSLs or complex config files—and run them anywhere. Prefect handles the heavy lifting for you out of the box: automatic state tracking, failure handling, real-time monitoring, and more.
Under the hood, Prefect manages scheduling and execution details so you can focus on writing clean, maintainable Python code instead of juggling infrastructure. Whether you’re a data engineer or part of a broader team, Prefect makes it simple to move from idea to production-ready workflows—fast.
Main features
| Feature | Description |
|---|---|
| Pure Python Development | Write workflows in native Python—no DSLs, YAML, or special syntax. Full support for type hints, async/await, and modern Python patterns. Use your existing IDE, debugger, and testing tools. |
| State & Recovery | Robust state management that tracks success, failure, and retry states. Resume interrupted runs from the last successful point, and cache expensive computations to avoid unnecessary rework. |
| Flexible & Portable Execution | Start flows locally for easy development, then deploy them anywhere—from a single process to containers, Kubernetes, or cloud services—without locking into a vendor. Infrastructure is defined by code (not just configuration), making it simple to scale or change environments. |
| Event-Driven | Trigger flows on schedules, external events, or via API. Pause flows for human intervention or approval. Chain flows together based on states, conditions, or any custom logic. |
| Dynamic Runtime | Create tasks dynamically at runtime based on actual data or conditions. Easily spawn new tasks and branches during execution for truly data-driven workflows. |
| Modern UI | Real-time flow run monitoring, logging, and state tracking through an intuitive interface. View dependency graphs and DAGs automatically—just run your flow and open the UI. |
| CI/CD First | Test and simulate flows like normal Python code, giving you fast feedback during development. Integrate seamlessly into your existing CI/CD pipeline for automated testing and deployment. |
| Actually open source | Host the Prefect orchestration server yourself using containers, Kubernetes, or infrastructure-as-code—or let us handle server operations with our fully managed cloud service. Keep your workflow infrastructure wherever it belongs while choosing the orchestration model that fits your needs. |
Get started
Quickstart
In five minutes, create your first deployable workflow tracked with Prefect.
Prefect Cloud
Supercharge Prefect with enhanced governance, security, and performance capabilities.
Upgrade to Prefect 3
Upgrade from Prefect 2 to Prefect 3 to get the latest features and performance enhancements.
Tutorials
Schedule a flow
Get a workflow off of your laptop and run it on remote infrastructure on a schedule.
Build a data pipeline
Build a resilient and performant data pipeline with Prefect’s primitives.
Extract data from websites
Handle data dependencies and use pagination to efficiently extract data from websites.
Set up a platform
Use Prefect as a platform for your teams’ data pipelines.
Debug a data pipeline
Learn how to debug flow runs that fail, crash, or hang.
Send alerts on failure
Set up an email alert to notify your teams about failures.
Start building
Develop
Write, run, configure, and observe workflows and their tasks, results, artifacts, and more.
Deploy
Run workflows in local processes or deploy them to dynamically provisioned infrastructure.
Automate
Enable workflows to react to their environment with events, automations, and webhooks.
Join our community
Join Prefect’s vibrant community of nearly 30,000 engineers to learn with others and share your knowledge!
Was this page helpful?