Hugging Face Scraper — Trending Models, Datasets & Spaces

Track the fastest-moving corner of AI. This Actor pulls trending and top models, datasets, and Spaces from the official Hugging Face Hub API — downloads, likes, trending scores, pipeline tags, and update timestamps in clean JSON/CSV.

What you can extract:

• 🔥 Trending models — what the AI community is adopting right now
• 📚 Datasets & Spaces — same metrics, same clean schema
• 🔍 Search & filters — by keyword, org (meta-llama, mistralai…), or tag
• 📈 Sort options — trending, downloads, likes, newest, recently updated

Quick start

1. Click Run — default fetches the top 50 trending models
2. Add a search query ("llama", "OCR", "text-to-speech") or an org name
3. Schedule daily + diff runs = a new-model alert system for your niche

Sample output

{
    "assetType": "model",
    "id": "meta-llama/Llama-4-70B-Instruct",
    "url": "https://huggingface.co/meta-llama/Llama-4-70B-Instruct",
    "author": "meta-llama",
    "pipelineTag": "text-generation",
    "libraryName": "transformers",
    "downloads": 2841022,
    "likes": 15320,
    "trendingScore": 312,
    "gated": true,
    "tags": ["text-generation", "llama", "conversational"],
    "createdAt": "2026-04-05T16:00:12.000Z",
    "lastModified": "2026-06-28T09:31:44.000Z"
}

Use cases

• AI newsletters & content — weekly "what's trending" data, automated
• ML engineering — monitor new models in your task category (e.g. ASR, OCR)
• VC / market research — org-level activity tracking across labs
• Competitive intel — watch competitors' orgs for new releases
• RAG/agents — MCP-ready: let your agent query the AI ecosystem live

Input reference

Field	Type	Notes
assetTypes	array	models, datasets, spaces
searchQueries	array	Keyword searches
authors	array	Org/user names
filterTags	array	HF tags like text-generation
sort	select	trending / downloads / likes / updated / newest
maxItemsPerQuery	integer	Cost cap

Pricing

Pay per result row delivered.

FAQ

Official data? Yes — the public Hugging Face Hub API, no key needed.

How fresh is "trending"? It's HF's live trending score at run time — schedule daily to build a time series.