On this page
CrewAI lets you define multi-agent “crews” with roles, goals, and tasks that collaborate to complete work; integrating CrewAI with ZenML wraps those crews in reproducible pipelines with artifact tracking, orchestration, and an easy path from local experiments to production.
Features with ZenML
- Pipeline orchestration. Run CrewAI crews as ZenML steps inside reproducible pipelines.
- Artifact management. Capture inputs and outputs for lineage, versioning, and auditability.
- Built-in evaluation hooks. Add post-run checks or eval steps to monitor quality over time.Infrastructure agnostic. Scale from local runs to Kubernetes, Airflow, and other ZenML stacks.
- Composable workflows. Combine crews with retrieval, evals, and deployment steps in one DAG.
Main Features
- Multi-agent crews. Define collaborators with roles, goals, and tools.
- Task delegation. Break work into tasks and route them to the right agent.
- Collaborative execution. Agents coordinate to research, draft, and refine outputs.
- Pluggable patterns. Start from examples like research and writing crews and customize.
How to use ZenML with CrewAI
from zenml import ExternalArtifact, pipeline, step
from crewai_agent import crew
@step
def run_crewai(query: str) -> str:
result = crew.kickoff(inputs={"city": query})
return str(result)
@pipeline
def crewai_travel_pipeline() -> str:
q = ExternalArtifact(value="Berlin")
return run_crewai(q.value)
if __name__ == "__main__":
print(crewai_travel_pipeline())
