LangChain Model Context Protocol (MCP) Adapters

Classes:

Name	Description
MultiServerMCPClient	Client for connecting to multiple MCP servers and loading LangChain-compatible tools, prompts and resources from them.

MultiServerMCPClient

Client for connecting to multiple MCP servers and loading LangChain-compatible tools, prompts and resources from them.

Methods:

Name	Description
__init__	Initialize a MultiServerMCPClient with MCP servers connections.
session	Connect to an MCP server and initialize a session.
get_tools	Get a list of all tools from all connected servers.
get_prompt	Get a prompt from a given MCP server.
get_resources	Get resources from a given MCP server.

__init__

__init__(
    connections: dict[str, Connection] | None = None,
) -> None

Initialize a MultiServerMCPClient with MCP servers connections.

Parameters:

Name	Type	Description	Default
connections	dict[str, Connection] | None	A dictionary mapping server names to connection configurations. If None, no initial connections are established.	None

Example: basic usage (starting a new session on each tool call)

from langchain_mcp_adapters.client import MultiServerMCPClient

client = MultiServerMCPClient(
    {
        "math": {
            "command": "python",
            # Make sure to update to the full absolute path to your math_server.py file
            "args": ["/path/to/math_server.py"],
            "transport": "stdio",
        },
        "weather": {
            # make sure you start your weather server on port 8000
            "url": "http://localhost:8000/mcp",
            "transport": "streamable_http",
        }
    }
)
all_tools = await client.get_tools()

Example: explicitly starting a session

from langchain_mcp_adapters.client import MultiServerMCPClient
from langchain_mcp_adapters.tools import load_mcp_tools

client = MultiServerMCPClient({...})
async with client.session("math") as session:
    tools = await load_mcp_tools(session)

session async

session(
    server_name: str, *, auto_initialize: bool = True
) -> AsyncIterator[ClientSession]

Connect to an MCP server and initialize a session.

Parameters:

Name	Type	Description	Default
server_name	str	Name to identify this server connection	required
auto_initialize	bool	Whether to automatically initialize the session	True

Raises:

Type	Description
ValueError	If the server name is not found in the connections

Yields:

Type	Description
AsyncIterator[ClientSession]	An initialized ClientSession

get_tools async

get_tools(
    *, server_name: str | None = None
) -> list[BaseTool]

Get a list of all tools from all connected servers.

Parameters:

Name	Type	Description	Default
server_name	str | None	Optional name of the server to get tools from. If None, all tools from all servers will be returned (default).	None

NOTE: a new session will be created for each tool call

Returns:

Type	Description
list[BaseTool]	A list of LangChain tools

get_prompt async

get_prompt(
    server_name: str,
    prompt_name: str,
    *,
    arguments: dict[str, Any] | None = None
) -> list[HumanMessage | AIMessage]

Get a prompt from a given MCP server.

get_resources async

get_resources(
    server_name: str, *, uris: str | list[str] | None = None
) -> list[Blob]

Get resources from a given MCP server.

Parameters:

Name	Type	Description	Default
server_name	str	Name of the server to get resources from	required
uris	str | list[str] | None	Optional resource URI or list of URIs to load. If not provided, all resources will be loaded.	None

Returns:

Type	Description
list[Blob]	A list of LangChain Blobs

Functions:

Name	Description
load_mcp_tools	Load all available MCP tools and convert them to LangChain tools.

load_mcp_tools async

load_mcp_tools(
    session: ClientSession | None,
    *,
    connection: Connection | None = None
) -> list[BaseTool]

Load all available MCP tools and convert them to LangChain tools.

Parameters:

Name	Type	Description	Default
session	ClientSession | None	MCP client session	required
connection	Connection | None	Optional connection config to use to create a new session if a session is not provided	None

Returns:

Type	Description
list[BaseTool]	a list of LangChain tools

Functions:

Name	Description
load_mcp_prompt	Load MCP prompt and convert to LangChain messages.

load_mcp_prompt async

load_mcp_prompt(
    session: ClientSession,
    name: str,
    *,
    arguments: dict[str, Any] | None = None
) -> list[HumanMessage | AIMessage]

Load MCP prompt and convert to LangChain messages.

Functions:

Name	Description
load_mcp_resources	Load MCP resources and convert them to LangChain Blobs.

load_mcp_resources async

load_mcp_resources(
    session: ClientSession,
    *,
    uris: str | list[str] | None = None
) -> list[Blob]

Load MCP resources and convert them to LangChain Blobs.

Parameters:

Name	Type	Description	Default
session	ClientSession	MCP client session	required
uris	str | list[str] | None	List of URIs to load. If None, all resources will be loaded. NOTE: if you specify None, dynamic resources will NOT be loaded, as they need the parameters to be provided, and are ignored by MCP SDK's session.list_resources() method.	None

Returns:

Type	Description
list[Blob]	A list of LangChain Blobs