Mar 6, 2026 10 min

    CrewAI Multi-Agent Systems: Build Collaborative AI Teams

    Create teams of specialized AI agents that collaborate on complex tasks — powered by Vincony's 400+ models for optimal role assignment.

    CrewAI Multi-Agent Agent Frameworks

    What Is CrewAI?

    CrewAI is a framework for orchestrating role-playing AI agents that work together. Each agent has a role, a goal, and a backstory — and they collaborate through structured delegation, much like a real team. Combined with Vincony, you can assign the best model to each agent role.

    Setting Up a Crew with Vincony

    from crewai import Agent, Task, Crew, Process
from langchain_openai import ChatOpenAI

# Use Vincony as the LLM provider for all agents
vincony_llm = ChatOpenAI(
    model="gpt-4.1",
    api_key="YOUR_VINCONY_API_KEY",
    base_url="https://api.vincony.com/v1"
)

# Define specialized agents
researcher = Agent(
    role="Senior Research Analyst",
    goal="Find and synthesize the latest information on {topic}",
    backstory="Expert at finding authoritative sources and extracting key insights.",
    llm=vincony_llm,
    verbose=True
)

writer = Agent(
    role="Technical Writer",
    goal="Create clear, engaging content from research findings",
    backstory="Skilled at translating complex topics into accessible prose.",
    llm=ChatOpenAI(
        model="claude-sonnet-4",
        api_key="YOUR_VINCONY_API_KEY",
        base_url="https://api.vincony.com/v1"
    )
)

editor = Agent(
    role="Editor-in-Chief",
    goal="Polish content for accuracy, clarity, and SEO",
    backstory="Meticulous editor with a keen eye for factual errors.",
    llm=ChatOpenAI(
        model="gemini-2.5-pro",
        api_key="YOUR_VINCONY_API_KEY",
        base_url="https://api.vincony.com/v1"
    )
)

    Defining Tasks & Workflows

    # Define tasks with dependencies
research_task = Task(
    description="Research the latest developments in {topic}. Find 5+ authoritative sources.",
    expected_output="A structured research brief with key findings and source citations.",
    agent=researcher
)

writing_task = Task(
    description="Write a 1500-word article based on the research brief.",
    expected_output="A polished article with introduction, body sections, and conclusion.",
    agent=writer,
    context=[research_task]  # Depends on research output
)

editing_task = Task(
    description="Review and edit the article for accuracy, clarity, and SEO optimization.",
    expected_output="Final publication-ready article with SEO metadata.",
    agent=editor,
    context=[writing_task]
)

# Assemble and run the crew
crew = Crew(
    agents=[researcher, writer, editor],
    tasks=[research_task, writing_task, editing_task],
    process=Process.sequential,
    verbose=True
)

result = crew.kickoff(inputs={"topic": "RAG vs Fine-Tuning in 2026"})

    Optimizing Model Assignment

    Not every agent needs the most powerful model. Use GPT-4.1 for complex reasoning, Claude for creative writing, Gemini Flash for quick classification tasks, and smaller models for formatting. Vincony's cost tracking shows exactly what each agent costs per run.

    Hierarchical Process

    For complex workflows, use CrewAI's hierarchical process where a manager agent delegates tasks dynamically based on the situation — re-assigning work if quality checks fail.

    Pricing

    CrewAI integration uses standard Vincony credit pricing per model call. A typical 3-agent crew run costs 5-15 credits depending on model selection and task complexity.

    Related Articles

    Related Tutorials & Resources

    Try It Free — 100 API Credits

    Start using these tools today with Vincony's free Developer plan.

    Get Free API Key