How to Use the Nuclino MCP in LangChain

Q: How does LangChain handle the hierarchical structure of Nuclino workspaces?

Your LangChain agent runs listteams to find your organization, then passes that ID to listworkspaces to locate specific areas. This sequential tool calling lets the agent navigate your entire Nuclino wiki hierarchy dynamically during a run.

Q: Can I trace Nuclino tool executions inside LangChain using LangSmith?

Yes, every call to tools like getitem or updateitem registers as a distinct step in your LangChain trace. You can inspect the exact markdown payloads and latency of your Nuclino wiki updates in real time.

Q: Can my LangChain agent map document relationships using this MCP Server?

Yes, your LangChain agent uses listcollections to trace the visual document relationship graph paths within a target Nuclino workspace. This allows the agent to understand how different wiki pages connect before it writes new content.

Feed your LangChain agents real-time wiki context and let them update Nuclino workspaces directly through deterministic tool chains.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect Nuclino MCP to LangChain

Create your Vinkius account to connect Nuclino to LangChain and route execution through our secure gateway. The platform manages server hosting, runtime updates, and security layers. Configuration requires no manual server provisioning.

GDPR Free for Subscribers

Setup Nuclino with LangChain

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Chain Nuclino discovery with LangChain agents

The `list_teams` tool acts as the entry point for your LangChain agent to discover root organizational units before traversing down the workspace hierarchy. By feeding the resulting team IDs directly into `list_workspaces`, your agent maps out the structure of your wiki dynamically without hardcoded IDs. This MCP Server eliminates manual configuration by letting the LangChain agent trace the organizational path from team to workspace in a single execution loop. You no longer need to hardcode IDs or manually map routes.

Trace Nuclino document edits in LangSmith

The `update_item` tool alters the sync tree immediately when your LangChain agent appends new wiki edits to existing pages. Every markdown update and payload modification is tracked in LangSmith, giving you full visibility into the exact token usage and latency of your document writes. If a LangChain agent attempts to overwrite a partial state in Nuclino, you see the input parameters and the raw markdown response in your tracing dashboard. This monitoring ensures you catch formatting errors or unwanted edits before they affect your live team documentation.

Build multi-step wiki research loops with this MCP Server

The `search_items` tool executes an indexed semantic search across your entire Nuclino team to find relevant document UUIDs for your LangChain agent. Your agent takes these search results and pipes them directly into `get_item` to pull down the raw markdown content for synthesis. Because LangChain manages state across tool calls, the agent can recursively search, read, and cross-reference multiple Nuclino pages. It builds a complete picture of your internal knowledge graph without manual prompt engineering.

Setup guide

Set up Nuclino MCP in LangChain

Prerequisites

Python 3.10+ installed
langchain-mcp-adapters + langgraph packages
Active Vinkius subscription with a valid endpoint token

1

Install dependencies
Run pip install langchain-mcp-adapters langgraph langchain-openai. The MCP adapters package converts MCP tools into native LangChain BaseTool objects.
2

Connect via HTTP transport
Use MultiServerMCPClient with "transport": "http" pointing to your Vinkius endpoint. Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com.
3

Create a ReAct agent
Pass the discovered tools to create_react_agent() from LangGraph. The agent automatically routes Nuclino tool calls through the MCP protocol.
4

Run with any LLM
Swap ChatOpenAI for ChatAnthropic, ChatGoogleGenerativeAI, or any LangChain-compatible model. The MCP tools work identically across all providers.

agent.py

from langchain_mcp_adapters.client import MultiServerMCPClient
from langgraph.prebuilt import create_react_agent
from langchain_openai import ChatOpenAI

async with MultiServerMCPClient({
    "nuclino-mcp": {
        "transport": "http",
        "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp",
    }
}) as client:
    tools = client.get_tools()

    agent = create_react_agent(
        ChatOpenAI(model="gpt-4o"),
        tools,
    )
    result = await agent.ainvoke({
        "messages": "List recent Nuclino transactions"
    })
    print(result["messages"][-1].content)

Get your connection token →

Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Nuclino. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.

Why Choose Vinkius

Vinkius connects your tools to AI with real-time monitoring and automatic cost savings — all from one dashboard.

Connect Nuclino now

Real-time monitoring

Live

visibility into every interaction

Connect your favorite tools to your AI and see exactly what's happening — every request, every response, in real time.

Built-in savings

60%

lower AI costs

Vinkius compresses data between your apps and your AI automatically. Lower bills every month — no configuration required.

Single dashboard

One

place for every integration

Every tool your AI connects to, managed from a single screen. One account, complete control.

Common questions about Nuclino MCP in LangChain

Your LangChain agent runs `list_teams` to find your organization, then passes that ID to `list_workspaces` to locate specific areas. This sequential tool calling lets the agent navigate your entire Nuclino wiki hierarchy dynamically during a run.

Yes, every call to tools like `get_item` or `update_item` registers as a distinct step in your LangChain trace. You can inspect the exact markdown payloads and latency of your Nuclino wiki updates in real time.

You should configure your LangChain agent's system prompt to require explicit human approval before invoking the Nuclino `delete_item` tool. Since deletion is irreversible, adding a confirmation step in your chain prevents accidental data loss.

Yes, your LangChain agent uses `list_collections` to trace the visual document relationship graph paths within a target Nuclino workspace. This allows the agent to understand how different wiki pages connect before it writes new content.

Your raw markdown payloads and workspace structures never pass through external intermediate servers. The Vinkius MCP Server sandbox isolates the connection, executing your queries locally and passing data directly to your local LangChain runtime.

Use it with your favorite AI tools

Connect this server to Cursor, Claude, VS Code, and more.

OpenAI Agents SDK sdk-python

Google ADK sdk-python

Pydantic AI sdk-python

Vercel AI SDK sdk-typescript