How to Use the ClickHouse (Vector Search) MCP in LangChain

Q: How do I configure the ClickHouse (Vector Search) MCP Server in a LangChain chain?

Install the adapter package and use MultiServerMCPClient pointing to your Vinkius endpoint. Pass the tools from client.gettools() directly to your agent executor.

Q: Does the ClickHouse (Vector Search) MCP Server support multi-server aggregation in LangChain?

Yes, the MultiServerMCPClient aggregates this server alongside other tools. Your agent can pull data from ClickHouse and immediately pipe it into a different API.

Q: How does the agent discover the database schema?

The agent runs listtables and describetable to read structural schemas. This lets it write valid SQL queries for executesql without hardcoded schema maps.

Get raw ClickHouse vector search speed inside your LangChain reasoning pipelines without writing manual drivers.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect ClickHouse (Vector Search) MCP to LangChain

Create your Vinkius account to connect ClickHouse (Vector Search) 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 ClickHouse (Vector Search) with LangChain

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Run ClickHouse MCP Server search in LangChain chains

The `vector_search` tool lets your LangChain agents query high-dimensional data directly. By exposing this function as a native tool, your agent runs mathematical distance queries across millions of rows and gets the closest matches in milliseconds. You don't have to configure database clients or write custom wrapper classes for your pipeline. The agent checks the distance traces to verify accuracy. If the results need context, the agent chains this step with `describe_table` to inspect active column schemas and ensure the data matches the expected embedding dimensions before running further calculations.

Execute raw SQL updates on active clusters

The `execute_sql` tool runs raw database queries directly from your agent. This lets your pipeline create temporary tables, update metadata, or aggregate search results on the fly without human intervention. You get full control over tables without leaving the LangChain framework. To keep the pipeline stable, the agent uses `list_databases` and `list_tables` to discover what is available before running queries. It prevents syntax errors and missing table exceptions by inspecting the environment dynamically during the run.

Monitor cluster health inside LangSmith

The `get_table_stats` tool pulls internal structural states to monitor active cluster health. This MCP Server exposes these metrics directly to your pipeline, letting LangSmith capture the outputs. You get clean observability into latency and cluster limits without setting up external dashboards. You can also query `get_version` to check for specific HNSW index support on the target instance. This ensures your LangChain application only triggers advanced vector distance algorithms when the underlying engine supports them.

Setup guide

Set up ClickHouse (Vector Search) 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 ClickHouse (Vector Search) 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({
    "clickhouse-vector-search-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 ClickHouse (Vector Search) 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 ClickHouse. 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 ClickHouse (Vector Search) 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 ClickHouse (Vector Search) MCP in LangChain

Install the adapter package and use `MultiServerMCPClient` pointing to your Vinkius endpoint. Pass the tools from `client.get_tools()` directly to your agent executor.

Every database query and vector search runs as a standard tool call. LangSmith automatically captures inputs, outputs, and execution latency for tools like `vector_search` and `execute_sql`.

Yes, the `MultiServerMCPClient` aggregates this server alongside other tools. Your agent can pull data from ClickHouse and immediately pipe it into a different API.

The agent runs `list_tables` and `describe_table` to read structural schemas. This lets it write valid SQL queries for `execute_sql` without hardcoded schema maps.

Vinkius runs the MCP Server in an isolated sandbox. Your database credentials never leak to the client, and the agent only sees the clean tool interfaces like `vector_search`.

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