Hugging Face Scraper: Models, Datasets, Downloads, No Login avatar

Hugging Face Scraper: Models, Datasets, Downloads, No Login

Pricing

from $0.70 / 1,000 results

Go to Apify Store
Hugging Face Scraper: Models, Datasets, Downloads, No Login

Hugging Face Scraper: Models, Datasets, Downloads, No Login

Hugging Face Hub models, datasets and Spaces: downloads, likes, tags, task, library, license and dates. Search by keyword, author, task, library or language. No API key. $1 per 1,000 results, platform usage included.

Pricing

from $0.70 / 1,000 results

Rating

0.0

(0)

Developer

One Dollar Store

One Dollar Store

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

12 hours ago

Last modified

Features

Limited permissions

Share

Hugging Face Scraper: models, datasets and Spaces with downloads and tags

$1 per 1,000 results. Platform usage included. No rental, no minimum.

Search the Hugging Face Hub by keyword, author, task, library, language or license and get one row per model, dataset or Space: downloads, likes, trending score, tags, license, dates and more. The data comes from the public Hub API. No Hugging Face account or API key is needed.

Open it in Apify Console

New to Apify? [Create a free account](https://www.apify.com?fpr=ymnoit) (referral link).

How much does Hugging Face Scraper cost?

ResultsPrice
1,000$1.00
10,000$10.00
100,000$100.00

One flat price, platform usage included, so there is no separate compute or proxy bill. A daily run of 100 models costs about $0.10 a day, or about $3 a month.

Why use it

  • Models, datasets and Spaces in one actor. Pick the kind of repository with one input field.
  • Filters that match how the Hub works. Task, library, language, license, author, any other Hub tag, minimum downloads and minimum likes.
  • Ready to use fields. License, base models, datasets a model was trained on, paper ids and parameter count are pulled out of the tags for you, next to the full tag list.
  • Cheap to monitor. Sort by trending score, last updated or newest and schedule a daily run to watch a task or an organization.

How to use Hugging Face Scraper

  1. Open the actor, type a search word, press Start.
  2. Or paste this input:
{ "resourceType": "models", "search": "llama", "sort": "downloads", "maxItems": 50 }
  1. Download the results as JSON, CSV or Excel, or read them through the API.

Use it from code

curl -X POST "https://api.apify.com/v2/acts/one_dollar_store~hugging-face-models-scraper/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
-H "Content-Type: application/json" \
-d '{"resourceType": "models", "task": "text-classification", "language": "de", "maxItems": 100}'
from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("one_dollar_store/hugging-face-models-scraper").call(
run_input={"resourceType": "models", "task": "text-classification", "language": "de", "maxItems": 100}
)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
print(item["downloads"], item["id"], item.get("license"))

Use it from an AI assistant

Claude, ChatGPT, Cursor and other assistants can call this actor through Apify's MCP server. Add it with this server address:

https://mcp.apify.com?tools=one_dollar_store/hugging-face-models-scraper

Then ask, for example: "List the 30 most downloaded Apache 2.0 text-generation models on Hugging Face."

Switching from another Hugging Face actor

Most Hugging Face actors take a search word and a result limit. Here they are search and maxItems, with resourceType to choose models, datasets or Spaces. The core output fields are id, downloads, likes, pipelineTag, tags and lastModified.

Input

FieldTypeDefaultWhat it does
resourceTypetextmodelsmodels, datasets or spaces.
searchtextnoneWords to find in the repository name or owner, for example llama.
authortextnoneOnly repositories owned by this user or organization, for example google.
tasktextnoneModels: a pipeline tag such as text-generation. Datasets: a task category such as question-answering. Not used for Spaces.
librarytextnoneModels: transformers, diffusers, gguf, peft and so on. Datasets: datasets, pandas, polars. Spaces: the SDK, for example gradio or streamlit.
languagetextnoneA language code such as en, de or zh. Not used for Spaces.
licensetextnoneA license id such as apache-2.0, mit or cc-by-4.0.
extraTagslist of textnoneAny other Hub tags. Every tag has to match.
sorttextdownloadsdownloads (last 30 days), likes, trendingScore, lastModified or createdAt. Highest or latest first.
minDownloadsnumbernoneSkip repositories with fewer downloads in the last 30 days.
minLikesnumbernoneSkip repositories with fewer likes.
maxItemsnumber100Stop after this many results. Minimum 10.

You can leave every filter empty. Then the actor lists the Hub in the order you chose. Example with filters:

{
"resourceType": "datasets",
"task": "question-answering",
"language": "en",
"license": "apache-2.0",
"sort": "likes",
"maxItems": 200
}

What data does Hugging Face Scraper return?

FieldMeaning
typemodel, dataset or space.
id, author, name, urlRepository id (owner/name), owner, short name and Hub page.
downloadsDownloads in the last 30 days, as the Hub counts them. Models and datasets.
downloadsAllTimeDownloads since the repository was created. Models and datasets.
likes, trendingScoreLikes and the Hub's trending score.
createdAt, lastModifiedDates in UTC.
gated, gatedModeWhether you must request access to the files, and auto or manual.
licenseLicense id from the tags.
tagsEvery Hub tag of the repository.
pipelineTag, libraryNameModels: the task and the main library.
parametersModels: parameter count, when the Hub reports it from safetensors files.
baseModels, trainedOnDatasets, arxivIdsModels: base models, training datasets and paper ids from the tags.
descriptionDatasets: the start of the dataset card, up to 600 characters.
taskCategories, languages, sizeCategory, modalities, formatsDatasets: details from the tags.
sdk, subdomain, stage, hardware, usesModels, usesDatasetsSpaces: SDK, app address, running state, hardware and the models and datasets it uses.

Fields that do not apply, or that the Hub has no value for, are left out of the row.

Sample results

Real rows from test runs on 9 October 2026 (tag lists and long descriptions shortened).

[
{"type": "model", "id": "meta-llama/Llama-3.1-8B-Instruct", "author": "meta-llama", "name": "Llama-3.1-8B-Instruct",
"url": "https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct", "downloads": 6191157, "downloadsAllTime": 188719078,
"likes": 8297, "trendingScore": 73, "createdAt": "2024-07-18T08:56:00.000Z", "lastModified": "2024-09-25T17:00:57.000Z",
"gated": true, "gatedMode": "manual", "license": "llama3.1", "tags": ["transformers", "safetensors", "llama", "text-generation"],
"pipelineTag": "text-generation", "libraryName": "transformers", "parameters": 8030261248,
"baseModels": ["meta-llama/Llama-3.1-8B"], "arxivIds": ["2204.05149"]},
{"type": "dataset", "id": "FreedomIntelligence/medical-o1-reasoning-SFT", "author": "FreedomIntelligence", "name": "medical-o1-reasoning-SFT",
"url": "https://huggingface.co/datasets/FreedomIntelligence/medical-o1-reasoning-SFT", "downloads": 21732, "downloadsAllTime": 246791,
"likes": 1191, "trendingScore": 3, "createdAt": "2024-12-28T03:29:08.000Z", "lastModified": "2025-04-22T15:11:21.000Z",
"gated": false, "license": "apache-2.0", "tags": ["task_categories:question-answering", "language:en", "language:zh"],
"description": "News [2025/04/22] We split the data and kept only the medical SFT dataset (medical_o1_sft.json)...",
"taskCategories": ["question-answering", "text-generation"], "languages": ["en", "zh"], "sizeCategory": "10K<n<100K",
"modalities": ["text"], "formats": ["json"], "arxivIds": ["2412.18925"]},
{"type": "space", "id": "ResembleAI/Chatterbox", "author": "ResembleAI", "name": "Chatterbox",
"url": "https://huggingface.co/spaces/ResembleAI/Chatterbox", "likes": 1756, "trendingScore": 0,
"createdAt": "2025-05-27T09:45:30.000Z", "lastModified": "2026-01-14T18:45:12.000Z", "gated": false,
"tags": ["gradio", "mcp-server", "region:us"], "sdk": "gradio", "subdomain": "resembleai-chatterbox",
"stage": "RUNNING", "hardware": "zero-a10g", "usesModels": ["ResembleAI/chatterbox"]}
]

Limits

  • Only public repositories. Private ones never appear. Gated models and datasets are listed with their details, but you need access on Hugging Face to download their files.
  • search matches the repository name and owner, not the text of the model card. To find a topic, use the task, library, language, license and extraTags filters.
  • A task or language filter matches Hub tags. A model can carry a task tag without it being the model's main task, so check pipelineTag if that matters.
  • Downloads and likes are the Hub's numbers on the day of the run. Spaces have no download counts.
  • No model card text (datasets get the first 600 characters), file lists, discussions or version history.
  • The Hub allows 500 API calls per 5 minutes for anonymous use from one address. Each call here carries up to 500 repositories, and the actor waits when the Hub asks it to slow down. A 1,500 result run took under two minutes in our test.

FAQ

How much does it cost to scrape 10,000 Hugging Face models? $10.

Do I pay for compute or proxy on top? No. Platform usage is included in the $1 per 1,000 price. The actor needs no proxy.

Where does the data come from? From the public list service of the Hugging Face Hub, the same data the Hub website shows to every visitor. No Hugging Face account is used.

Is it legal? The actor reads public data only, with no login. Counts, tags and dates are facts the Hub shows to everyone. Each model and dataset has its own license, shown in the license field: follow it when you use the model or data itself, and follow Hugging Face's terms and the law where you live.

Is this an official Hugging Face product? No. It is an independent tool and is not affiliated with or endorsed by Hugging Face.

How do I watch for new models on a task? Schedule a daily run with task set, sort set to createdAt and a small maxItems, then drop the id values you already have.

Feedback and support

Found a problem or want a field added? Open an issue on the Issues tab of this actor and it gets read. If the actor saved you time, an honest rating on the Reviews tab helps other people find it.

Do I need an Apify account? Yes. Actors run from an Apify account, and a free account is enough to start: it includes $5 of platform usage every month. You can sign up here. This is a referral link and it costs you nothing.

More from The One Dollar Store

Every actor here costs $1 per 1,000 results, platform usage included. See the whole store: https://apify.com/one_dollar_store

Domains and apps

Changelog

  • 26 October 2026: clearer description and headings, field details for AI tools. Price and output unchanged.
  • October 2026: first release.