Integration

Catalogian + CrewAI

CrewAI enables you to build teams of AI agents that collaborate on complex tasks. By integrating Catalogian's MCP tools, your CrewAI agents get live access to product catalog data — enabling crews that monitor feed changes, analyze pricing trends, and generate reports, all backed by real-time product data.

Connect CrewAI to Catalogian's MCP server and your agents can browse product schemas, filter catalog rows, query change history, and download data. A 'catalog analyst' agent can check what changed today, while a 'pricing strategist' agent recommends actions — each using Catalogian tools as part of their task workflow.

The integration requires no custom API wrappers. CrewAI's tool system connects to Catalogian's MCP endpoint directly. Define your crew's tasks around product data questions, and Catalogian's tools provide the answers your agents need to complete their work.

How it works

01

Configure Catalogian as a tool source

Point CrewAI at your Catalogian MCP endpoint (POST /v1/mcp) with your API key. The 14 MCP tools become available to any agent in your crew.

02

Define agents with catalog tasks

Create CrewAI agents with roles like 'catalog monitor' or 'pricing analyst'. Assign tasks that involve querying product data, tracking changes, or generating reports.

03

Run your crew on live data

When the crew executes, agents call Catalogian tools to get real product data. No mock data or static files — every tool call returns live catalog information.

CrewAI agent with Catalogian MCP tools

from crewai import Agent, Task, Crew
from crewai_tools import MCPServerAdapter

catalogian = MCPServerAdapter(
    server_url="https://catalogian.com/v1/mcp",
    headers={"x-api-key": "cat_sk_live_..."}
)

analyst = Agent(
    role="Catalog Analyst",
    goal="Monitor product feed changes and flag anomalies",
    tools=catalogian.tools
)

task = Task(
    description="Check my-catalog for price changes in the last 24 hours. "
                "Flag any price drops greater than 20%.",
    agent=analyst
)

crew = Crew(agents=[analyst], tasks=[task])
result = crew.kickoff()

Frequently asked questions

What can a CrewAI agent do with Catalogian?

Anything the MCP tools support: browse catalog schemas, filter and search product rows, query delta changes, download snapshots, and check source health. Agents use these tools to answer questions about your product data.

Can I build a multi-agent crew around Catalogian data?

Yes. For example, one agent monitors for changes, another analyzes the impact, and a third generates a report. All three can access Catalogian tools independently, or pass data between them through CrewAI's task handoff system.

Does the integration require a specific CrewAI version?

MCP tool support in CrewAI is available through the crewai-tools package. Use the latest version for full MCP server adapter compatibility.

Start monitoring your product feeds

Free plan includes 1 source and 50,000 SKUs. No credit card required.

Start monitoring free