Firecrawl is an enterprise-grade web scraping platform that turns any website into clean, AI-ready data. Combined with Groq's fast inference through MCP, you can build intelligent agents that scrape websites, extract structured data, and research the web with natural language instructions.
Key Features:
pip install openai python-dotenvexport GROQ_API_KEY="your-groq-api-key"
export FIRECRAWL_API_KEY="your-firecrawl-api-key"import os
from openai import OpenAI
from openai.types import responses as openai_responses
client = OpenAI(
base_url="https://api.groq.com/openai/v1",
api_key=os.getenv("GROQ_API_KEY")
)
tools = [
openai_responses.tool_param.Mcp(
server_label="firecrawl",
server_url="https://mcp.firecrawl.dev/v2/mcp",
headers={"Authorization": f"Bearer {os.getenv('FIRECRAWL_API_KEY')}"},
type="mcp",
require_approval="never",
)
]
response = client.responses.create(
model="openai/gpt-oss-120b",
input="Scrape https://console.groq.com/docs/models and provide an overview of available models",
tools=tools,
temperature=0.1,
top_p=0.4,
)
print(response.output_text)Extract data in specific JSON formats across multiple sources:
response = client.responses.create(
model="openai/gpt-oss-120b",
input="""Extract pricing from https://openai.com, https://anthropic.com, https://groq.com
Return JSON:
{
"company_name": "string",
"pricing_plans": [{"plan_name": "string", "price": "string", "features": ["string"]}]
}""",
tools=tools,
temperature=0.1,
)
print(response.output_text)Search the web and synthesize findings across multiple pages:
response = client.responses.create(
model="openai/gpt-oss-120b",
input="""Research "latest trends in AI model inference speed and performance":
1. Recent developments (2024-2025)
2. Key companies and technologies
3. Performance benchmarks
4. Future trends
Provide a comprehensive report with citations.""",
tools=tools,
temperature=0.1,
)
print(response.output_text)Map a site, then scrape the pages that matter:
response = client.responses.create(
model="openai/gpt-oss-120b",
input="""Map https://console.groq.com/docs, pick the three pages about tool use,
scrape them, and summarize the key findings from each.""",
tools=tools,
temperature=0.1,
)
print(response.output_text)Firecrawl MCP provides these core tools for web scraping, data extraction, and research. See the Firecrawl MCP tools reference for the full, current list.
| Tool | Description |
|---|---|
firecrawl_scrape | Scrape a single URL as Markdown, or as structured JSON with a schema |
firecrawl_map | List the URLs on a site without fetching their content |
firecrawl_search | Search the web and optionally extract content from search results |
firecrawl_crawl | Crawl many pages under a site and return the results |
firecrawl_check_crawl_status | Check a crawl job by ID |
firecrawl_parse | Parse documents such as PDFs and spreadsheets |
firecrawl_agent | Start an autonomous research job that returns structured data |
firecrawl_agent_status | Check an agent job and retrieve its results |
Use allowed_tools on the MCP tool to limit the model to the tools your application needs.
Challenge: Build an AI-powered competitive intelligence system that monitors competitor websites, extracts key business metrics, and generates automated reports using Firecrawl and Groq!
For more detailed documentation and resources on building web intelligence applications with Groq and Firecrawl, see: