Skip to main content
Workflows enable deterministic, multi-step AI pipelines. Unlike teams where agents decide dynamically, workflows follow a pre-defined sequence.

​
What is a workflow?

A workflow is a structured sequence of steps:
  • Each step is an agent, team, or function
  • Steps execute in order
  • Output from one step can be input to the next
  • Support for conditionals, loops, and parallel execution

​
Quick example

from agno.workflow import Workflow, Step

researcher = Agent(name="Researcher", tools=[DuckDuckGoTools()])
writer = Agent(name="Writer")
editor = Agent(name="Editor")

workflow = Workflow(
    steps=[
        Step(agent=researcher, name="Research"),
        Step(agent=writer, name="Write"),
        Step(agent=editor, name="Edit")
    ]
)

result = workflow.run("Write an article about AI safety")

​
When to use workflows

  • Content pipelines: Research → Write → Edit → Publish
  • Data processing: Extract → Transform → Load
  • Review processes: Create → Review → Approve
See cookbook examples for complete patterns.

Build docs developers (and LLMs) love

Get started for freeTalk to us