How to Use the DVC MCP in LangChain

Q: How do I track DVC experiment history inside a LangChain agent?

Use listexperiments within your agent's toolset. The agent calls this tool to retrieve past run parameters, which you can then pass to other tools in your chain.

Q: Can LangChain run parallel queries across multiple DVC projects?

Yes. By combining listprojects with a MultiServerMCPClient, your agent can query several repositories simultaneously. It aggregates the project metadata back into a single chain context.

Q: How does LangSmith trace DVC tool operations?

LangSmith automatically captures every call to getproject or listviews made by the agent through the MCP Server. You get full visibility into the exact parameters, execution time, and model decisions for each run.

Q: What happens if an agent requests a view that does not exist?

The getview tool returns an error message. Your agent reads this error and can call listviews to find a valid alternative.

Run multi-step ML experiment audits directly inside your LangChain chains.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect DVC MCP to LangChain

Create your Vinkius account to connect DVC 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 DVC with LangChain

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Chain pipeline outputs to DVC project states

Feed your model training outputs directly into your LangChain runs. Your agent calls `list_projects` to find the active workspace, then pulls the exact configuration using `get_project`. This lets you pass real-time ML architecture states directly into the next step of your chain without hardcoded scripts. By using this MCP Server, the output of one step naturally feeds the next. Your agent handles the logic, checking project files and feeding the parameters straight into your evaluation chains.

Audit experiment histories via the MCP Server

Stop guessing which run produced a model. Your agent runs `list_experiments` to pull the complete history of your training runs. Because this runs inside LangChain, every single tool call is traced in LangSmith, giving you a clean audit log of inputs, outputs, and latency. You see the exact parameters your agent inspected. If a run looks off, the agent calls `get_view` to inspect the specific dataset slice used during that run. No more untraceable background runs.

Dynamic dataset view extraction

Let your agent pull specific data subsets on the fly. By calling `list_views` and `get_view`, the agent inspects the exact data partitions defined in your DVC setup. This keeps your training pipelines grounded. The agent confirms the dataset version before starting a new run, making sure your team never trains on outdated data slices again.

Setup guide

Set up DVC 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 DVC 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({
    "dvc-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 DVC 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 DVC. 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 DVC 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 DVC MCP in LangChain

Use `list_experiments` within your agent's toolset. The agent calls this tool to retrieve past run parameters, which you can then pass to other tools in your chain.

Yes. By combining `list_projects` with a MultiServerMCPClient, your agent can query several repositories simultaneously. It aggregates the project metadata back into a single chain context.

LangSmith automatically captures every call to `get_project` or `list_views` made by the agent through the MCP Server. You get full visibility into the exact parameters, execution time, and model decisions for each run.

The `get_view` tool returns an error message. Your agent reads this error and can call `list_views` to find a valid alternative.

All calls run inside a secure, ephemeral V8 isolate sandbox on Vinkius. Your experiment parameters, user profiles, and dataset views are never stored or exposed to external networks.

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