Hugging Face MCP

How to Install and Use Hugging Face MCP

Hugging Face MCP (Managed Connector Protocol) is a tool that helps connect your Large Language Model (LLM) to the Hugging Face Hub and many AI applications like Gradio. Below, you will find clear steps on how to install and use Hugging Face MCP in different environments.

Installing Hugging Face MCP in Claude Desktop or claude.ai

To add Hugging Face MCP as a connector in Claude Desktop or claude.ai, simply add it to your account.

Here is what you do next:

1. Go to https://claude.ai/settings/connectors.
2. Look for "Hugging Face" in the gallery.
3. Click to add it to your account.

Alternatively, you can quickly add it using this link:
Add Hugging Face Connector

Installing Hugging Face MCP in Claude Code

If you want to use Hugging Face MCP in Claude Code, the installation involves running a command in your terminal.

Follow these steps:

1. Run the command below to add Hugging Face MCP:

claude mcp add hf-mcp-server -t http https://huggingface.co/mcp?login

1. Start claude and follow the instructions to finish authentication.

If you prefer to use an authentication token, use this command instead:

claude mcp add hf-mcp-server -t http https://huggingface.co/mcp -H "Authorization: Bearer <YOUR_HF_TOKEN>"

Replace <YOUR_HF_TOKEN> with your actual Hugging Face token.

Installing Hugging Face MCP in Gemini CLI

Gemini CLI users can also add Hugging Face MCP easily. Here’s how:

1. Run this command to add Hugging Face MCP:

gemini mcp add -t http huggingface https://huggingface.co/mcp?login

1. Start gemini and then run:

/mcp auth huggingface

to authenticate the extension.

For a smoother experience, you can install the full Hugging Face Gemini CLI extension including MCP Server support:

gemini extensions install https://github.com/huggingface/hf-mcp-server

Installing Hugging Face MCP in VSCode

If you want to integrate Hugging Face MCP with Visual Studio Code (VSCode), you can do this by adding it from the MCP gallery:

• Visit https://code.visualstudio.com/mcp and add the Hugging Face connector.

If you want to configure it manually, add the following to your mcp.json configuration file:

"huggingface": {
    "url": "https://huggingface.co/mcp",
    "headers": {
        "Authorization": "Bearer <YOUR_HF_TOKEN>"
    }
}

Replace <YOUR_HF_TOKEN> with your Hugging Face token.

Running Hugging Face MCP Server Locally

You can also run the Hugging Face MCP server on your own computer:

1. Using npx:

npx @llmindset/hf-mcp-server             # Start in STDIO mode
npx @llmindset/hf-mcp-server-http        # Start in Streamable HTTP mode
npx @llmindset/hf-mcp-server-json        # Start in Streamable HTTP (JSON RPC) mode

1. Using Docker:

First, pull the latest image:

docker pull ghcr.io/evalstate/hf-mcp-server:latest

Then, run the server with this command:

docker run --rm -p 3000:3000 ghcr.io/evalstate/hf-mcp-server:latest

This will start the server and open the management web interface at http://localhost:3000/.

---

Using these easy steps, you can install and start using Hugging Face MCP in your preferred environment quickly and easily. Just follow the commands closely, and you'll have your MCP Server connected and running.