Connect AgentDeals MCP tools to CrewAI agents via native mcps field. Multi-agent pricing research workflows with search, compare, and track capabilities.
CrewAI is a popular framework for orchestrating multi-agent AI workflows. It has native MCP support via the mcps field on agents, making it easy to give your crew access to AgentDeals tools.
AgentDeals provides 4 MCP tools with data on 1,580+ developer deals across 65 categories: search_deals, compare_vendors, track_changes, and plan_stack.
pip install crewai
from crewai import Agent, Task, Crew
researcher = Agent(
role="Developer Tools Researcher",
goal="Find the best free developer tools and compare pricing",
backstory="You are an expert at evaluating developer tool "
"free tiers and identifying the best options.",
mcps=[
{
"url": "https://agentdeals.dev/mcp",
}
],
)
task = Task(
description="Search for free database hosting options and "
"compare the top 3 by free tier limits, risk level, "
"and pricing at scale.",
expected_output="A comparison table of the top 3 free databases "
"with recommendations.",
agent=researcher,
)
crew = Crew(agents=[researcher], tasks=[task])
result = crew.kickoff()
print(result)
CrewAI's native mcps field connects directly to the AgentDeals remote server. No additional packages or local installation needed.
For stdio transport (local), use:
mcps=[
{
"command": "npx",
"args": ["-y", "agentdeals"],
}
]Create a research crew with a searcher and an analyst working together to evaluate developer tools.
searcher = Agent(
role="Tool Searcher",
goal="Find free developer tools across categories",
backstory="Expert at discovering developer tool free tiers.",
mcps=[{"url": "https://agentdeals.dev/mcp"}],
)
analyst = Agent(
role="Pricing Analyst",
goal="Compare vendor pricing and assess risk",
backstory="Expert at evaluating pricing sustainability.",
mcps=[{"url": "https://agentdeals.dev/mcp"}],
)
search_task = Task(
description="Find the top 5 free monitoring tools. "
"Include data ingestion limits and retention.",
expected_output="List of 5 monitoring tools with free tier details.",
agent=searcher,
)
analysis_task = Task(
description="Compare the monitoring tools found above. "
"Assess free tier risk and recommend the best option "
"for a startup with 10 microservices.",
expected_output="Comparison table with risk assessment "
"and final recommendation.",
agent=analyst,
context=[search_task],
)
crew = Crew(agents=[searcher, analyst], tasks=[search_task, analysis_task])
result = crew.kickoff()
print(result)Set up a crew that monitors pricing changes and generates alerts for your stack.
monitor = Agent(
role="Pricing Monitor",
goal="Track developer tool pricing changes",
backstory="You watch for free tier changes that could "
"affect development teams.",
mcps=[{"url": "https://agentdeals.dev/mcp"}],
)
monitor_task = Task(
description="Check for recent pricing changes across "
"hosting, database, and auth categories. Flag any "
"free tier removals or significant limit reductions.",
expected_output="Summary of recent changes with risk assessment.",
agent=monitor,
)
crew = Crew(agents=[monitor], tasks=[monitor_task])
result = crew.kickoff()
print(result)AgentDeals exposes 4 MCP tools. Your CrewAI agent can call any of them:
search_dealsSearch 1,580+ deals by keyword, category, or vendor
compare_vendorsSide-by-side comparison of free tiers, limits, and risk
track_changesTrack pricing changes, removals, and new deals
plan_stackStack recommendations with cost projections at scale
Full tool documentation: /api/docs · Setup guide for more clients
Use AgentDeals tools directly in your AI editor. Works with Claude Desktop, Cursor, Cline, Windsurf, and any MCP-compatible client.
claude mcp add agentdeals -- npx -y agentdealsWorks with Claude Desktop, Cursor, Cline, Windsurf → Full setup guide
Connect AgentDeals to LangChain agents via langchain-mcp-adapters — Python code examples and multi-agent workflows
Connect AgentDeals to n8n via MCP Server Trigger node — no-code pricing monitoring and vendor comparison workflows
Connect AgentDeals to Vercel AI SDK via experimental_createMCPClient() — React/Next.js code examples