mcp-use

Backed by

Build and deploy MCP agents

Spin-up and aggregate MCP servers through a single endpoint and zero friction.

Or use our open source libraries

pip install mcp-use

npm install mcp-use

Github

Google Drive

Mem0AI

Your MCP Server

Slack

Library used by devs from

Vercel for MCP

1-click deploy MCP servers and aggregate them through a single endpoint

# Connect to remote server pool

client = MCPClient(server_pool="marketing_servers")

# Use any model you like

llm = AnyModelProvider(model="any-4o", api_key=...)

# Create agent with tools

agent = MCPAgent(llm=llm, client=client, max_steps=30)

# Run the query

result = await agent.run("Create a post about our latest release")

class MCPClient:

def __init__(self):

# Initialize session and client objects

self.session: Optional[ClientSession] = None

self.exit_stack = AsyncExitStack()

self.anthropic = Anthropic()

# methods will go here

async def connect_to_server(self, server_script_path: str):

is_python = server_script_path.endswith('.py')

is_js = server_script_path.endswith('.js')

if not (is_python or is_js):

raise ValueError("Server script must be a .py or .js file")

command = "python" if is_python else "node"

server_params = StdioServerParameters(

command=command,

args=[server_script_path],

env=None

)

stdio_transport = await self.exit_stack.enter_async_context(stdio_client(server_params))

self.stdio, self.write = stdio_transport

self.session = await self.exit_stack.enter_async_context(ClientSession(self.stdio, self.write))

await self.session.initialize()

# List available tools

response = await self.session.list_tools()

tools = response.tools

print("\nConnected to server with tools:", [tool.name for tool in tools])

async def process_query(self, query: str) -> str:

"""Process a query using Claude and available tools"""

messages = [

{

"role": "user",

"content": query

}

]

response = await self.session.list_tools()

available_tools = [{

"name": tool.name,

"description": tool.description,

"input_schema": tool.inputSchema

} for tool in response.tools]

# Initial Claude API call

response = self.anthropic.messages.create(

model="claude-3-5-sonnet-20241022",

max_tokens=1000,

messages=messages,

tools=available_tools

)

# Process response and handle tool calls

final_text = []

assistant_message_content = []

for content in response.content:

if content.type == 'text':

final_text.append(content.text)

assistant_message_content.append(content)

elif content.type == 'tool_use':

tool_name = content.name

tool_args = content.input

# Execute tool call

result = await self.session.call_tool(tool_name, tool_args)

final_text.append(f"[Calling tool {tool_name} with args {tool_args}]")

assistant_message_content.append(content)

messages.append({

"role": "assistant",

"content": assistant_message_content

})

messages.append({

"role": "user",

"content": [

{

"type": "tool_result",

"tool_use_id": content.id,

"content": result.content

}

]

})

# Get next response from Claude

response = self.anthropic.messages.create(

model="claude-3-5-sonnet-20241022",

max_tokens=1000,

messages=messages,

tools=available_tools

)

final_text.append(response.content[0].text)

return "

".join(final_text)

MCP Agent & Client SDK

MCP agent in 3 lines

Install the OSS mcp-use library and turn any MCP call into an AI agent. Just pass the an ID or URL and the SDK pulls MCP servers config and streams results out of the box.

Auth

Security

Servers

Pool

mcp-use gateway

MCP servers routing

Cache

Memory

Monitor

Metrics

MCP Gateway

One endpoint for every MCP server

The managed MCP gateway routes, authenticates, and load-balances all your servers. Hosted, ephemeral, or on-prem. Built-in OAuth, ACLs, metrics, and tracing give you production-grade observability from day one.

Stdio MCP servers

Hosted short-lived MCP servers

Third-party MCP servers

External remote services

Cloud

Your MCP servers

Managed or self-hosted

mcp-use hosting

MCP Servers

Spin up, plug in, or bring your own

Deploy fully managed MCP servers in our cloud, sandbox local VMs on demand, or proxy third-party servers behind the gateway, all managed in one dashboard.

The control plane for MCP

All your MCP infrastructure in one place

Clients

Claude Desktop

Your AI product

MCP client