Pass work from one agent to the next when each step depends on the previous output.

from praisonaiagents import Agent, AgentFlow

researcher = Agent(name="Researcher", instructions="Research topics")
analyst = Agent(name="Analyst", instructions="Analyze research findings")
writer = Agent(name="Writer", instructions="Write based on analysis")
flow = AgentFlow(steps=[researcher, analyst, writer])
flow.start("AI trends in 2024")

The user chains agents in order; each step receives the prior output until the final result is ready.

Sequential Workflow

Pass work from one agent to the next, like an assembly line.

How Context Flows

Each agent receives the previous output automatically.

When to Use

Code

from praisonaiagents import Agent, AgentFlow

researcher = Agent(name="Researcher", instructions="Research topics")
analyst = Agent(name="Analyst", instructions="Analyze research findings")
writer = Agent(name="Writer", instructions="Write based on analysis")

# Sequential: researcher → analyst → writer
flow = AgentFlow(steps=[researcher, analyst, writer])
result = flow.start("AI trends in 2024")

Example: Blog Post Pipeline

Step	Agent	Receives	Outputs
1	Researcher	Topic	Facts, data
2	Analyst	Facts	Key insights
3	Writer	Insights	Draft article
4	Editor	Draft	Polished post

With Callbacks

from praisonaiagents import AgentFlow, WorkflowHooksConfig

flow = AgentFlow(
    steps=[researcher, analyst, writer],
    hooks=WorkflowHooksConfig(
        on_step_complete=lambda name, r: print(f"✅ {name} done")
    )
)

Parallel

All at once (faster)

Routing

Send to specialists

Parallel Workflow

Orchestrator Pattern

​Sequential Workflow

​How Context Flows

​When to Use

​Code

​Example: Blog Post Pipeline

​With Callbacks

​Related