Introduction

What is a workflow?

A workflow is an event-driven, step-based way to control the execution flow of an application.

Your application is divided into sections called steps. A step receives an event, does some work, and returns another event. That returned event triggers the next step whose type annotation accepts it.

That is the whole model. A step can call an LLM, run retrieval, ask for human input, update shared state, or dispatch a batch of work. The event types describe the edges of the workflow, and regular Python describes the logic inside each edge.

Why workflows?

As generative AI applications become more complex, it becomes harder to manage the flow of data and control the execution of the application. Workflows provide a way to manage this complexity by breaking the application into smaller, more manageable pieces.

Other frameworks and LlamaIndex itself have attempted to solve this problem previously with directed acyclic graphs (DAGs) but these have a number of limitations that workflows do not:

• Logic like loops and branches needed to be encoded into the edges of graphs, which made them hard to read and understand.
• Passing data between nodes in a DAG created complexity around optional and default values and which parameters should be passed.
• DAGs did not feel natural to developers trying to develop complex, looping, branching AI applications.

The event-based pattern and plain Python approach of Workflows resolves these problems.

Branches are ordinary if statements that return different event types. Loops are steps that return an event handled by an earlier step. Concurrent work is a step that returns list[Event], paired with another step that accepts list[Event]. When the flow needs to become dynamic, you can send events directly from the Context.

Note

The Workflows library can be installed standalone, via pip install llama-index-workflows. However, llama-index-core comes with an installation of Workflows included.

In order to maintain the llama_index API stable and avoid breaking changes, when installing llama-index-core or the llama-index umbrella package, Workflows can be accessed with the import path llama_index.core.workflow.

Getting Started

Tip

Workflows make async a first-class citizen, and this page assumes you are running in an async environment. What this means for you is setting up your code for async properly. If you are already running in a server like FastAPI, or in a notebook, you can freely use await already!

If you are running your own Python scripts, it’s best practice to have a single async entry point.

async def main():
    w = MyWorkflow(...)
    result = await w.run(...)
    print(result)


if __name__ == "__main__":
    import asyncio

    asyncio.run(main())

Here is the smallest useful shape: generate something, pass it to another step, then stop with a result.

from workflows import Workflow, step
from workflows.events import (
    Event,
    StartEvent,
    StopEvent,
)

# `pip install llama-index-llms-openai` if you don't already have it
from llama_index.llms.openai import OpenAI


class JokeEvent(Event):
    joke: str


class JokeFlow(Workflow):
    llm = OpenAI(model="gpt-4.1")

    @step
    async def generate_joke(self, ev: StartEvent) -> JokeEvent:
        topic = ev.topic

        prompt = f"Write your best joke about {topic}."
        response = await self.llm.acomplete(prompt)
        return JokeEvent(joke=str(response))

    @step
    async def critique_joke(self, ev: JokeEvent) -> StopEvent:
        joke = ev.joke

        prompt = f"Give a thorough analysis and critique of the following joke: {joke}"
        response = await self.llm.acomplete(prompt)
        return StopEvent(result=str(response))


w = JokeFlow(timeout=60, verbose=False)
result = await w.run(topic="pirates")
print(str(result))

There are a few moving pieces here, so let’s go through them one at a time.

Defining Workflow Events

class JokeEvent(Event):
    joke: str

Events are user-defined Pydantic objects. You control the attributes and any other auxiliary methods. In this case, our workflow relies on a single user-defined event, the JokeEvent.

Setting up the Workflow Class

class JokeFlow(Workflow):
    llm = OpenAI(model="gpt-4.1")
    ...

Our workflow is implemented by subclassing the Workflow class. For simplicity, we attached a static OpenAI LLM instance.

Workflow Entry Points

class JokeFlow(Workflow):
    ...

    @step
    async def generate_joke(self, ev: StartEvent) -> JokeEvent:
        topic = ev.topic

        prompt = f"Write your best joke about {topic}."
        response = await self.llm.acomplete(prompt)
        return JokeEvent(joke=str(response))

    ...

Here, we come to the entry-point of our workflow. While most events are user-defined, there are two special-case events, the StartEvent and the StopEvent that the framework provides out of the box. Here, the StartEvent signifies where to send the initial workflow input.

The StartEvent is a bit of a special object since it can hold arbitrary attributes. Here, we accessed the topic with ev.topic, which would raise an error if it wasn’t there. You could also do ev.get("topic") to handle the case where the attribute might not be there without raising an error.

For further type safety, you can also subclass the StartEvent.

At this point, you may have noticed that we haven’t explicitly told the workflow what events are handled by which steps. Instead, the @step decorator is used to infer the input and output types of each step. Furthermore, these inferred input and output types are also used to verify for you that the workflow is valid before running!

Workflow Exit Points

class JokeFlow(Workflow):
    ...

    @step
    async def critique_joke(self, ev: JokeEvent) -> StopEvent:
        joke = ev.joke

        prompt = f"Give a thorough analysis and critique of the following joke: {joke}"
        response = await self.llm.acomplete(prompt)
        return StopEvent(result=str(response))

    ...

Here, we have our second, and last step, in the workflow. We know it’s the last step because the special StopEvent is returned. When the workflow encounters a returned StopEvent, it immediately stops the workflow and returns whatever we passed in the result parameter.

In this case, the result is a string, but it could be a dictionary, list, or any other object.

You can also subclass the StopEvent class for further type safety.

Running the Workflow

w = JokeFlow(timeout=60, verbose=False)
result = await w.run(topic="pirates")
print(str(result))

Lastly, we create and run the workflow. There are some settings like timeouts (in seconds) and verbosity to help with debugging.

The .run() method starts the workflow and returns a WorkflowHandler. The handler is awaitable, so await w.run(...) waits for the final result. If you need streamed events while the workflow is running, keep the handler:

handler = w.run(topic="pirates")
async for ev in handler.stream_events():
    ...
result = await handler

The keyword arguments passed to run() become fields of the special StartEvent that is automatically emitted to start the workflow. As we have seen, in this case topic is accessed from the step with ev.topic.

How to choose the right workflow API

Most workflow code should use typed step inputs and typed returns. That gives validation, diagrams, and readable code:

@step
async def retrieve(self, ev: StartEvent) -> Retrieved:
    ...

@step
async def synthesize(self, ev: Retrieved) -> StopEvent:
    ...

Use the other APIs when the shape of the work asks for them:

Use this	When
Return `A	B`
Return list[A]	A step has a finite batch and can produce all work items before downstream workers start.
Accept list[A]	A step needs the full batch of results before continuing.
ctx.send_event(...)	A step needs to emit events incrementally, emit an unknown number of events, or send an event from outside the workflow while it is running.
ctx.collect_events(...)	You used ctx.send_event and need to manually wait for a known set of events.
ctx.store	Steps need shared per-run state.
Resource(...)	Steps need clients, indexes, models, config, or other dependencies that should not live in serialized state.

The type-first APIs are easier to validate and visualize. The context APIs are more flexible, but they make you own more of the bookkeeping.

Validation

Before a workflow runs, Workflows validates the event graph described by your step signatures. It checks that start and stop events are present, produced events have consumers, consumed events have producers, and the graph does not contain accidental dead ends.

Most validation failures are useful design feedback. They usually mean one of these is true:

Symptom	Usual cause
A step consumes an event that is never produced	A return annotation is missing, or the event is only sent dynamically with ctx.send_event.
A step produces an event that nobody consumes	The next step has the wrong event type, or the branch is unfinished.
The workflow has no terminal event	No reachable step returns StopEvent or a custom StopEvent subclass.

For intentionally dynamic workflows, keep as much as possible in the type annotations and use ctx.send_event for the timing. If a step is intentionally unreachable from static analysis, skip that specific check on the step:

@step(skip_graph_checks=["reachability"])
async def receive_webhook(self, ev: WebhookEvent) -> StopEvent:
    ...

You can also call workflow.validate() directly in tests or startup code. Resource config files are validated by default; resource factories are only resolved if you call validate(validate_resources=True).

Examples

To help you become more familiar with the workflow concept and its features, LlamaIndex documentation offers example notebooks that you can run for hands-on learning:

• Common Workflow Patterns walks you through common usage patterns like looping and state management using simple workflows. It’s usually a great place to start.
• RAG + Reranking shows how to implement a real-world use case with a fairly simple workflow that performs both ingestion and querying.
• Citation Query Engine similar to RAG + Reranking, the notebook focuses on how to implement intermediate steps in between retrieval and generation. A good example of how to use the Context object in a workflow.
• Corrective RAG adds some more complexity on top of a RAG workflow, showcasing how to query a web search engine after an evaluation step.
• Utilizing Concurrency explains how to manage the parallel execution of steps in a workflow, something that’s important to know as your workflows grow in complexity.

RAG applications are easy to understand and offer a great opportunity to learn the basics of workflows. However, more complex agentic scenarios involving tool calling, memory, and routing are where workflows excel.

The examples below highlight some of these use-cases.

• ReAct Agent is obviously the perfect example to show how to implement tools in a workflow.
• Function Calling Agent is a great example of how to use the LlamaIndex framework primitives in a workflow, keeping it small and tidy even in complex scenarios like function calling.
• CodeAct Agent is a great example of how to create a CodeAct Agent from scratch.
• Human In The Loop: Story Crafting is a powerful example showing how workflow runs can be interactive and stateful. In this case, to collect input from a human.
• Reliable Structured Generation shows how to implement loops in a workflow, in this case to improve structured output through reflection.
• Query Planning with Workflows is an example of a workflow that plans a query by breaking it down into smaller items, and executing those smaller items. It highlights how to stream events from a workflow, execute steps in parallel, and looping until a condition is met.
• Writing Durable Workflows shows how to checkpoint workflow context and resume a run after a restart.

Last but not least, a few more advanced use cases that demonstrate how workflows can be extremely handy if you need to quickly implement prototypes, for example from literature:

• Advanced Text-to-SQL
• JSON Query Engine
• Long RAG
• Multi-Step Query Engine
• Multi-Strategy Workflow
• Router Query Engine
• Self Discover Workflow
• Sub-Question Query Engine