MCP

Access and Analyze Hugging Face Datasets

MCP server that connects to Hugging Face Dataset Viewer API for browsing, searching, filtering, and analyzing datasets hosted on Hugging Face Hub.

Connect

Works with huggingface

김정석

Maintainer?

Spark score

out of 100

Updated 4 months ago

Version 1.0.0

Models

claude

Add to Favorites

Why it matters

Effortlessly interact with Hugging Face datasets. View, analyze, search, and filter datasets hosted on the Hugging Face Hub using a dedicated MCP server.

Outcomes

What it gets done

Validate dataset existence and accessibility

Retrieve and paginate dataset content

Search and filter dataset rows using SQL-like conditions

Download datasets in Parquet format

Install

Add it to your toolbox

Run in your project directory:

curl -fsSL https://spark.entire.vc/get/vb-dataset-viewer | bash

Capabilities

Tools your agent gets

validate

Checks whether a dataset exists and is accessible

get_info

Retrieves detailed information about a dataset

get_rows

Retrieves dataset content with pagination

get_first_rows

Retrieves the first rows from a dataset split

get_statistics

Retrieves statistics for a dataset split

search_dataset

Searches for text in a dataset

filter

Filters rows using SQL-like conditions

get_parquet

Downloads the full dataset in Parquet format

Overview

Dataset Viewer MCP Server

What it does

This MCP server provides access to the Hugging Face Dataset Viewer API, enabling dataset exploration and analysis through tools for validation, information retrieval, content access, statistics, searching, filtering, and Parquet downloads.

How it connects

Use this connector when you need to work with datasets hosted on Hugging Face Hub, whether public or private, and want to inspect metadata, retrieve specific rows, search content, apply filters, or download complete datasets in Parquet format.

Source README

MCP server for working with the Hugging Face Dataset Viewer API, providing capabilities to view, analyze, search, and filter datasets hosted on Hugging Face Hub.

Installation

From Source

git clone https://github.com/privetin/dataset-viewer.git
cd dataset-viewer
uv venv
source .venv/bin/activate
uv add -e .

Configuration

Claude Desktop

{
  "mcpServers": {
    "dataset-viewer": {
      "command": "uv",
      "args": [
        "--directory",
        "parent_to_repo/dataset-viewer",
        "run",
        "dataset-viewer"
      ]
    }
  }
}

Available Tools

Tool	Description
`validate`	Checks whether a dataset exists and is accessible
`get_info`	Retrieves detailed information about a dataset
`get_rows`	Retrieves dataset content with pagination
`get_first_rows`	Retrieves the first rows from a dataset split
`get_statistics`	Retrieves statistics for a dataset split
`search_dataset`	Searches for text in a dataset
`filter`	Filters rows using SQL-like conditions
`get_parquet`	Downloads the full dataset in Parquet format

Features

Uses the dataset:// URI scheme for accessing Hugging Face datasets
Supports dataset configurations and splits
Provides paginated access to dataset content
Handles authentication for private datasets
Supports searching and filtering dataset content
Provides dataset statistics and analysis

Environment Variables

Optional

HUGGINGFACE_TOKEN - Your Hugging Face API token for accessing private datasets

Notes

Requires Python 3.12 or higher and the uv package installer. MIT License.

Discussion