Skip to main content
huggingface.health
Datasets

Is Hugging Face Datasets Server down? Live status

Open datasets + streaming loader

Live status for Hugging Face Datasets — the Hub datasets API and datasets-server preview/query service.

Datasets API Status

Live, in-browser checks against Datasets's public APIs. Refreshes every 15s.

Checking…

Datasets

Hub datasets API and Datasets Server

3 endpoints
  • Datasets Directory
    huggingface.co/datasets
    —ms
    warming up
  • Datasets API
    Public datasets listing endpoint
    —ms
    warming up
  • Datasets Server
    datasets-server.huggingface.co
    —ms
    warming up
Polling every 15s · Last updated — · Next in 15sOpen Datasets ↗

The Datasets Server powers row previews and SQL queries in dataset viewers. When it's degraded, the raw dataset files on the Hub are still downloadable.

Sponsored · Datasets
Ask Data
Hub · Streaming

I'm Data — your guide to Hugging Face Datasets. Ask about loading, streaming, or the Datasets Server.

Browse Datasets

Sponsored content. huggingface.health may earn a referral reward when you sign up via this widget. Answers are informational only — not professional advice.

About Datasets

Hugging Face Datasets hosts 250k+ open datasets with a unified loading API, streaming support, and a Datasets Server that provides row-level previews and DuckDB-style queries in the browser.

Datasets at a glance

Datasets on Hub
250k+
Library
🤗 datasets (Apache-2.0)
Streaming
Yes (`streaming=True`)
Viewer
Datasets Server + DuckDB
Formats
Parquet, JSON, CSV, Arrow, WebDataset

Common Datasets issues

  • Viewer not available

    Datasets Server needs precomputed splits. If your dataset is too large, private, or uses a custom loading script, the viewer may skip it — files still load fine.

  • load_dataset hangs on download

    Large multi-shard datasets download serially by default. Use `num_proc=` for parallel shard downloads, or switch to `streaming=True` for on-demand fetch.

  • Auth error on gated dataset

    Gated datasets require acceptance of terms plus an HF_TOKEN with read access. Run `huggingface-cli login` before `load_dataset()`.

Background

The 🤗 datasets library launched in 2020 as a unified way to load and stream NLP corpora. The Datasets Server followed to power in-browser previews and SQL over Parquet, making the Hub the default place to publish open datasets across modalities.

Frequently asked

Why is the dataset viewer 'Not available'?
Datasets Server processes each dataset asynchronously. Very large or private datasets may not have precomputed splits. The raw files remain downloadable with `datasets.load_dataset()`.
Can I stream a dataset without downloading it?
Yes — pass `streaming=True` to `load_dataset()`. The datasets library fetches shards on demand, which is essential for TB-scale corpora.

About this check

All data is provided on a best-effort basis with no guarantees about accuracy or availability. This is an unofficial status page and is not affiliated with, endorsed by, or sponsored by Hugging Face, Inc.

Checks run client-side from your browser against Datasets's public endpoints. Latency reflects your network path, not the server's health alone. See our resources for more on how to read these checks.

Related on this site