CrewAI

CrewAI is a framework for orchestrating role-playing AI agents. SideSeat captures runs (traces) from crews, agents, and tasks.

Prerequisites

• SideSeat running locally (sideseat)
• Python 3.9+
• Model/provider credentials configured

Quick Start

1. Start SideSeat

   npx sideseat

2. Install dependencies

   pip

   pip install crewai "sideseat[crewai]"

   uv

   uv add crewai "sideseat[crewai]"

3. Add telemetry

   from sideseat import SideSeat, Frameworks
   from crewai import Agent, Task, Crew
   
   SideSeat(framework=Frameworks.CrewAI)
   
   researcher = Agent(
       role='Researcher',
       goal='Find information',
       backstory='Expert researcher'
   )
   
   task = Task(
       description='Research AI trends',
       expected_output='Summary of trends',
       agent=researcher
   )
   
   crew = Crew(agents=[researcher], tasks=[task])
   result = crew.kickoff()

4. View runs

   Open http://localhost:5388 to see your runs.

Without SideSeat SDK

1. Start SideSeat

   npx sideseat

2. Set the endpoint

   export OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:5388/otel/default

3. Install dependencies

   pip

   pip install crewai openinference-instrumentation-crewai opentelemetry-exporter-otlp

   uv

   uv add crewai openinference-instrumentation-crewai opentelemetry-exporter-otlp

4. Add telemetry

   from opentelemetry import trace
   from opentelemetry.sdk.trace import TracerProvider
   from opentelemetry.sdk.trace.export import BatchSpanProcessor
   from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
   from openinference.instrumentation.crewai import CrewAIInstrumentor
   
   provider = TracerProvider()
   provider.add_span_processor(BatchSpanProcessor(OTLPSpanExporter()))
   trace.set_tracer_provider(provider)
   CrewAIInstrumentor().instrument(tracer_provider=provider, skip_dep_check=True)

5. View runs

   Open http://localhost:5388 to see your runs.

Multi-Agent Crews

CrewAI’s multi-agent workflows are fully traced:

from crewai import Agent, Task, Crew, Process

# Define agents
researcher = Agent(
    role='Senior Researcher',
    goal='Uncover groundbreaking technologies',
    backstory='Expert at finding new tech trends'
)

writer = Agent(
    role='Tech Writer',
    goal='Create engaging content',
    backstory='Skilled at explaining complex topics'
)

# Define tasks
research_task = Task(
    description='Research the latest AI developments',
    expected_output='Detailed research report',
    agent=researcher
)

write_task = Task(
    description='Write a blog post about the research',
    expected_output='Engaging blog post',
    agent=writer
)

# Create crew
crew = Crew(
    agents=[researcher, writer],
    tasks=[research_task, write_task],
    process=Process.sequential
)

result = crew.kickoff()

SideSeat shows:

• Parent span for the crew execution
• Child spans for each agent’s work
• Child spans for each task execution

Extracted Attributes

Attribute	Description
crewai.crew.name	Crew identifier
crewai.agent.role	Agent role
crewai.task.description	Task being executed
crewai.process.type	sequential or hierarchical

Tools

Agent tools are automatically traced:

from crewai.tools import tool

@tool
def search_web(query: str) -> str:
    """Search the web for information."""
    return f"Results for {query}"

agent = Agent(
    role='Researcher',
    goal='Find information',
    backstory='Expert researcher',
    tools=[search_web]
)

Process Types

Both sequential and hierarchical processes are captured:

# Sequential (default)
crew = Crew(
    agents=[agent1, agent2],
    tasks=[task1, task2],
    process=Process.sequential
)

# Hierarchical
crew = Crew(
    agents=[manager, worker1, worker2],
    tasks=[task1, task2],
    process=Process.hierarchical,
    manager_llm=ChatOpenAI(model="gpt-5-mini")
)

Next Steps

• Python SDK — SDK reference
• Core Concepts — understanding runs, steps, and messages