LangChain

Use AllMCP tools in LangChain agents via langchain-mcp-adapters.

---

Install

install

pip install langchain-mcp-adapters langchain-anthropic langgraph

---

Minimal example

agent.py

import asyncio
from langchain_mcp_adapters.client import MultiServerMCPClient
from langchain_anthropic import ChatAnthropic
from langgraph.prebuilt import create_react_agent

async def main():
    client = MultiServerMCPClient({
        "allmcp": {
            "url": "https://go.allmcp.co/mcp/",
            "transport": "streamable_http",
            "headers": {"X-API-Key": "YOUR_API_KEY"},
        }
    })
    tools = await client.get_tools()
    model = ChatAnthropic(model="claude-sonnet-4-6")
    agent = create_react_agent(model, tools)

    result = await agent.ainvoke({
        "messages": [{"role": "user", "content": "List my Bitrix24 contacts"}]
    })
    print(result["messages"][-1].content)

asyncio.run(main())

Note

Auth header. Pass your key as X-API-Key: allmcp_xxx. If your transport layer strips custom headers, use Authorization: Bearer allmcp_xxx instead — both are accepted.

---

Multi-user setup

Note

One agent, many users. Append ?user_id= to the endpoint URL so each request acts as that user and uses the providers they’ve connected.

agent.py

async def run_for_user(user_id: str, prompt: str) -> str:
    client = MultiServerMCPClient({
        "allmcp": {
            "url": f"https://go.allmcp.co/mcp/?user_id={user_id}",
            "transport": "streamable_http",
            "headers": {"X-API-Key": "YOUR_API_KEY"},
        }
    })
    tools = await client.get_tools()
    agent = create_react_agent(ChatAnthropic(model="claude-sonnet-4-6"), tools)
    result = await agent.ainvoke({"messages": [{"role": "user", "content": prompt}]})
    return result["messages"][-1].content

---

With memory

agent.py

from langgraph.checkpoint.memory import MemorySaver

client = MultiServerMCPClient({...})
tools = await client.get_tools()
agent = create_react_agent(
    ChatAnthropic(model="claude-sonnet-4-6"),
    tools,
    checkpointer=MemorySaver(),
)

config = {"configurable": {"thread_id": "session-1"}}

# Turn 1
await agent.ainvoke({"messages": [{"role": "user", "content": "Connect Bitrix24"}]}, config)
# Turn 2 — agent remembers the previous turn
result = await agent.ainvoke({"messages": [{"role": "user", "content": "Now list my deals"}]}, config)

---

Using OpenAI instead of Anthropic

agent.py

from langchain_openai import ChatOpenAI

model = ChatOpenAI(model="gpt-4o")
agent = create_react_agent(model, tools)

Tip

Any LLM works. AllMCP tools work with any LangChain-compatible LLM.

---

See also

Frameworks overview

Browse providers See every provider you can connect and the tools each one offers.