Skip to main content

Quickstart

Run your first Flyte workflow in under 5 minutes — locally or on a cluster.

Key Concepts

Understand tasks, workflows, launch plans, and how Flyte executes pipelines.

User Guide

Learn Flyte’s Python SDK: data types, advanced composition, and productionizing.

Deployment

Deploy Flyte on AWS, GCP, or on-premises with Helm charts and Kubernetes.

​
What is Flyte?

Flyte is an open-source, Kubernetes-native workflow orchestration platform for building production-grade data and ML pipelines. It is used by teams at LinkedIn, Spotify, Freenome, Pachama, Warner Bros., and many others.

Strongly Typed

Every task input and output is typed. Flyte validates data at workflow boundaries, preventing silent failures.

Reproducible

Immutable executions and full data lineage mean you can always reproduce any run — exactly.

Kubernetes-Native

Flyte runs on any Kubernetes cluster. Scale from a local sandbox to thousands of parallel tasks on AWS, GCP, or Azure.

Multi-Language

Write in Python, Java, Scala, or any language via raw containers. First-class Python SDK with flytekit.

Dynamic Workflows

Build workflows that change shape at runtime based on data, branching, and conditional logic.

Plugin Ecosystem

Native integrations with Spark, Ray, PyTorch, TensorFlow, Airflow, and more through the plugin system.

​
Core architecture

Flyte is a multi-component platform. Each component is independently deployable and Helm-packaged:
ComponentRole
FlyteAdminControl plane. Stores workflow definitions, launches executions, exposes gRPC + REST API.
FlytePropellerExecution engine. Kubernetes operator that reconciles workflow states as a CRD.
FlyteIDLProtobuf-based interface definition language. The contract between all Flyte components.
DataCatalogArtifact caching layer. Enables task output memoization across runs.
FlyteConsoleWeb UI for monitoring workflows, executions, and artifacts.
flytectlCLI for managing all Flyte resources and running local sandboxes.

​
Get started in 3 steps

1

Install flytekit

Install Flyte’s Python SDK with pip:
pip install flytekit
2

Write a workflow

Create a file hello_world.py:
hello_world.py
from flytekit import task, workflow

@task
def say_hello(name: str) -> str:
    return f"Hello, {name}!"

@workflow
def hello_world_wf(name: str = "world") -> str:
    return say_hello(name=name)
3

Run it

Execute locally with:
pyflyte run hello_world.py hello_world_wf --name "Flyte"
Or on a remote cluster:
pyflyte run --remote hello_world.py hello_world_wf --name "Flyte"
Looking for Flyte 2? The new Python SDK is available at flyte-sdk. The documentation below covers Flyte 1.x.

​
Explore the docs

Core Concepts

Tasks, workflows, launch plans, projects, and domains explained.

Data Types & I/O

FlyteFile, FlyteDirectory, StructuredDataset, and typed interfaces.

Advanced Composition

Dynamic workflows, map tasks, subworkflows, conditionals, and checkpointing.

Flyte Agents

Extend Flyte with custom async or sync agents for external services.

Deployment Guide

Helm-based deployment for local, cloud, and on-premises environments.

CLI Reference

Full reference for flytectl — create, manage, and monitor everything.

Build docs developers (and LLMs) love

Get started for freeTalk to us