Dataset Viewer
Free浏览和分析 Hugging Face 数据集，支持搜索、过滤、统计和导出。
FreeFree tier
About Dataset Viewer
Dataset Viewer is an MCP (Model Context Protocol) server that integrates with the Hugging Face Dataset Viewer API to provide programmatic access for browsing, analyzing, and managing datasets hosted on the Hugging Face Hub. It supports resources via a dataset:// URI scheme, authentication for private datasets, paginated access, and offers a suite of tools for validation, metadata retrieval, row viewing, statistics, text search, SQL-like filtering, and full dataset export in Parquet format. Designed for developers and data scientists working with Hugging Face datasets in AI/ML workflows, it can be configured as a tool for Claude Desktop and other MCP-compatible clients.
Key Features
Integrates with Hugging Face Dataset Viewer API via MCP protocol
Resources with dataset:// URI scheme for accessing datasets
Supports dataset configurations and splits
Paginated access to dataset contents
Authentication support for private datasets via Hugging Face token
Tools: validate, get_info, get_rows, get_first_rows, get_statistics, search_dataset, filter, get_parquet
SQL-like filtering with WHERE and ORDER BY clauses
Full dataset export in Parquet format
Designed for use with Claude Desktop and other MCP clients
Pros & Cons
Pros
- Open source under MIT license
- Leverages MCP standard for easy integration with AI assistants
- Supports private datasets via token-based auth
- Provides both browsing and analytical capabilities (statistics, search, filter)
- Supports pagination for large datasets
- Includes Parquet export for efficient data handling
- Comprehensive set of tools covering validation, info, rows, statistics, search, filter
Cons
- Requires Python 3.12 or higher
- Dependency on the uv package manager for installation
- Limited to datasets on the Hugging Face Hub
- Requires manual setup and configuration for MCP clients like Claude Desktop
- No built-in graphical interface; command-line and API-based only
Best For
Browsing and inspecting Hugging Face datasets programmaticallyRetrieving dataset metadata and statistics for analysisSearching and filtering dataset contents for specific entriesAccessing private datasets with authenticationExporting datasets in Parquet format for local processingIntegrating dataset exploration into AI/ML workflows via MCP