How to Use the Hugging Face MCP in LangChain

Q: How do I connect Hugging Face to LangChain?

Install langchain-mcp-adapters and langgraph. Pass the Hugging Face endpoint to MultiServerMCPClient and use client.gettools() to arm your ReAct agent.

Q: Can LangChain agents read Hugging Face datasets?

They read the metadata and file structures. The agent uses listdatasetfiles to map out directories like train.parquet before deciding how to parse them.

Q: Does LangSmith trace these Hugging Face tool calls?

Yes. Every time your LangChain agent triggers getspace or getcollection, LangSmith logs the exact inputs, latency, and returned JSON payload.

Wire Hugging Face directly into your LangChain agents. Build multi-step ReAct pipelines that search models, inspect files, and track datasets.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect Hugging Face MCP to LangChain

Create your Vinkius account to connect Hugging Face 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 Hugging Face with LangChain

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Chain Hugging Face searches with this MCP Server

LangChain agents execute multi-step logic. You don't just query Hugging Face once. Your ReAct agent calls `list_models` via this MCP integration to find text-generation artifacts, then loops through the results using `get_model_tags` to filter out incompatible frameworks. That output feeds directly into the next chain link. The agent isolates the exact model ID and pipes it into `get_model` to pull author details and download counts. LangSmith tracks every token and API call across the entire sequence.

Inspect datasets before downloading

Blindly downloading massive datasets breaks pipelines. LangChain intercepts this problem by giving your agent the `list_datasets` tool to search for specific data structures. Once the agent finds a match, it doesn't just guess the format. It fires `list_dataset_files` to read the repository directory tree. The agent checks if the data lives in parquet files or raw CSVs before writing a single line of local extraction code. You get the exact file path without wasting bandwidth.

Automate model evaluations and bug reports

Models fail in production. When your LangChain evaluation pipeline detects a hallucination, it can automatically check if others hit the same wall. The agent runs `list_model_discussions` to scan the Hub for open bug reports on that specific model ID. If the issue is new, the agent takes action. It uses `create_discussion` to open a fresh thread on the repository, logging the failure trace directly to the author's page. You automate the entire feedback loop from error to open ticket.

Setup guide

Set up Hugging Face 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 Hugging Face 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({
    "hugging-face-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 Hugging Face 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 Hugging Face. 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 Hugging Face 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 Hugging Face MCP in LangChain

Install `langchain-mcp-adapters` and `langgraph`. Pass the Hugging Face endpoint to `MultiServerMCPClient` and use `client.get_tools()` to arm your ReAct agent.

They read the metadata and file structures. The agent uses `list_dataset_files` to map out directories like `train.parquet` before deciding how to parse them.

Yes. Every time your LangChain agent triggers `get_space` or `get_collection`, LangSmith logs the exact inputs, latency, and returned JSON payload.

Direct requests require hardcoding endpoints and parsing logic. The MCP standard exposes native tools like `list_spaces` that LangChain instantly converts into callable Python functions.

The framework passes a single endpoint token to the Vinkius V8 Isolate Sandbox. Your agent reads public repository metadata and user profile details via `get_user`, but the server destroys the execution context the millisecond the request finishes.

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