AutoGen

Integrate Firecrawl, the context API for AI agents, with Microsoft AutoGen to give multi-agent conversations live web search and scrape tools. Microsoft has archived AutoGen, so start new projects on Microsoft Agent Framework.

Microsoft has archived AutoGen and no longer maintains it. For new projects, Microsoft recommends Microsoft Agent Framework, and existing projects can follow the migration guide. The Firecrawl functions below are plain Python, so you can reuse them as tools there.

Setup

pip install -U "autogen-agentchat" "autogen-ext[openai]" firecrawl

Set your keys:

export FIRECRAWL_API_KEY=fc-YOUR-API-KEY
export OPENAI_API_KEY=sk-YOUR-OPENAI-KEY

Firecrawl as an AutoGen Tool

This example wraps Firecrawl’s scrape and search as AutoGen function tools, then lets a single AssistantAgent use them to answer a question.

import asyncio
from firecrawl import Firecrawl
from autogen_agentchat.agents import AssistantAgent
from autogen_agentchat.ui import Console
from autogen_ext.models.openai import OpenAIChatCompletionClient

firecrawl = Firecrawl()


def scrape_url(url: str) -> str:
    """Scrape a URL and return clean markdown."""
    result = firecrawl.scrape(url, formats=["markdown"])
    return result.markdown or ""


def web_search(query: str, limit: int = 5) -> list[dict]:
    """Search the web and return the top results."""
    result = firecrawl.search(query, limit=limit)
    return [
        {"title": r.title, "url": r.url, "snippet": r.description}
        for r in result.web or []
    ]


async def main() -> None:
    model = OpenAIChatCompletionClient(model="gpt-4o-mini")

    researcher = AssistantAgent(
        name="researcher",
        model_client=model,
        tools=[scrape_url, web_search],
        system_message=(
            "You are a web researcher. Use web_search to find candidate sources, "
            "then scrape_url to read the most relevant ones. Cite URLs in your answer."
        ),
    )

    await Console(
        researcher.run_stream(
            task="What does Firecrawl's /agent endpoint do? Cite the docs."
        )
    )


if __name__ == "__main__":
    asyncio.run(main())

Run it:

python researcher.py

Multi-Agent: Researcher + Writer

Hand Firecrawl output from a researcher agent to a writer agent in a round-robin team.

import asyncio
from firecrawl import Firecrawl
from autogen_agentchat.agents import AssistantAgent
from autogen_agentchat.conditions import MaxMessageTermination
from autogen_agentchat.teams import RoundRobinGroupChat
from autogen_agentchat.ui import Console
from autogen_ext.models.openai import OpenAIChatCompletionClient

firecrawl = Firecrawl()


def scrape_url(url: str) -> str:
    result = firecrawl.scrape(url, formats=["markdown"])
    return result.markdown or ""


def web_search(query: str, limit: int = 5) -> list[dict]:
    result = firecrawl.search(query, limit=limit)
    return [
        {"title": r.title, "url": r.url, "snippet": r.description}
        for r in result.web or []
    ]


async def main() -> None:
    model = OpenAIChatCompletionClient(model="gpt-4o-mini")

    researcher = AssistantAgent(
        name="researcher",
        model_client=model,
        tools=[scrape_url, web_search],
        system_message="Gather sources with web_search + scrape_url. Reply with bullet-point findings and URLs.",
    )

    writer = AssistantAgent(
        name="writer",
        model_client=model,
        system_message="Turn the researcher's findings into a 200-word briefing with inline citations.",
    )

    team = RoundRobinGroupChat(
        [researcher, writer],
        termination_condition=MaxMessageTermination(max_messages=6),
    )

    await Console(team.run_stream(task="Write a briefing on Firecrawl's crawl endpoint."))


if __name__ == "__main__":
    asyncio.run(main())

Notes

These examples use the synchronous Firecrawl client, which works inside AutoGen’s event loop for small workloads. For heavy concurrent scraping, use AsyncFirecrawl with async def tools, or use batch scrape.
Replace OpenAIChatCompletionClient with any AutoGen-supported model client (Azure OpenAI, Anthropic via autogen-ext, Ollama, etc.). Firecrawl is model-agnostic.
See the AutoGen docs for agent patterns beyond round-robin (selector, swarm, nested teams).

Getting Started

AI Tools

Agent Harnesses

MCP Clients and Gateways

LLM SDKs and Frameworks

Agentic Debugging

Cookbooks

Setup

Firecrawl as an AutoGen Tool

Multi-Agent: Researcher + Writer

Notes

​Setup

​Firecrawl as an AutoGen Tool

​Multi-Agent: Researcher + Writer

​Notes

Setup

Firecrawl as an AutoGen Tool

Multi-Agent: Researcher + Writer

Notes