Hugging Face Scraper Goat avatar

Hugging Face Scraper Goat

Pricing

from $3.00 / 1,000 results

Go to Apify Store
Hugging Face Scraper Goat

Hugging Face Scraper Goat

Scrape the Hugging Face Hub without a login. Search or list AI models, datasets, and spaces, sorted by downloads, likes, trending, or recency, and optionally filtered by task, library, or author. Returns clean rows with downloads, likes, task, library, tags, file count, and dates.

Pricing

from $3.00 / 1,000 results

Rating

0.0

(0)

Developer

Goutam Soni

Goutam Soni

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

6 hours ago

Last modified

Share

Hugging Face Scraper

Search and list AI models, datasets and spaces from the Hugging Face Hub at scale, no login or API key. Sort by downloads, likes, trending, or recency, and filter by task, library, or author. Every row comes back in one clean, structured shape with downloads, likes, task, library, tags, file count and dates.

What it does

  • Search - models, datasets or spaces matching your terms.
  • Top listings - leave the search empty to pull the top items by your chosen sort (for example the most-downloaded text-generation models).
  • Filters - narrow by task (e.g. text-generation), library (e.g. transformers), or author / organization (e.g. google).
  • Direct lookup - fetch specific Hub IDs.

Common uses: AI/ML market and trend research, model discovery and monitoring, dataset sourcing, competitor tracking, and building an AI catalog.

Input

FieldTypeDescription
resourcestringmodels, datasets, or spaces.
searchQueriesarrayTerms to search. Empty = top listing by sort.
directIdsarraySpecific Hub IDs, e.g. openai-community/gpt2.
sortstringdownloads, likes, trending, lastModified, createdAt.
filterstringOptional task or library filter.
authorstringOptional author / organization.
maxItemsPerSourceintegerCap per query. Default 50, up to 1000.
proxyConfigurationobjectOptional. Works without a proxy.

Example input

{
"resource": "models",
"searchQueries": ["llama", "whisper"],
"sort": "downloads",
"filter": "text-generation",
"maxItemsPerSource": 100
}

Output

Each item is one normalized model, dataset, or space.

{
"resource": "models",
"id": "openai-community/gpt2",
"author": "openai-community",
"name": "gpt2",
"downloads": 14024247,
"downloadsAllTime": null,
"likes": 3355,
"trendingScore": null,
"task": "text-generation",
"library": "transformers",
"tags": ["transformers", "pytorch", "text-generation"],
"gated": false,
"private": false,
"disabled": false,
"fileCount": 26,
"createdAt": "2022-03-02T23:29:04.000Z",
"lastModified": "2024-02-19T10:57:45.000Z",
"sha": "607a30d783dfa663caf39e06633721c8d4cfcd7e",
"url": "https://huggingface.co/openai-community/gpt2",
"scrapedAt": "2026-07-18T13:00:00.000Z"
}

Fields that do not apply (for example task on a dataset) come back as null, so the shape is always the same.

Notes

  • No login and no API key. Pick a resource, set your terms or a sort, and run.
  • Leave searchQueries empty and set sort to trending to track what is hot on the Hub right now.
  • maxItemsPerSource goes up to 1000 per query for large catalogs.

Found this useful? Leave a quick review. It takes a few seconds and it genuinely helps other people find the actor.