Hugging Face Models Scraper — Search, Downloads, Likes, Tags

Search the Hugging Face Hub for machine learning models by task, library, or keyword and export downloads, likes, license, and tags to a clean dataset. No login required.

What you get

• Model ID, author, model name, and direct Hugging Face URL for every result
• Pipeline task tag (text-generation, text-to-image, summarization, etc.) and library name (transformers, diffusers, sentence-transformers, peft, timm…)
• All Hub tags — useful for filtering by license, language, base model, dataset, or framework
• 30-day download count, likes count, and trending score
• Gated / private flags so you can spot restricted models
• Created and last-modified timestamps for freshness tracking
• Direct export to JSON, CSV, Excel, or Google Sheets

Use cases

• ML / model market research — measure popularity and adoption of open-source models
• Competitor tracking — monitor downloads and likes for models in your space (LLMs, image generation, embeddings, etc.)
• Build a model catalog — power an internal tool, blog post, or product comparison with up-to-date Hub metadata
• Lead generation — discover authors and organizations publishing in a specific task or library
• License audit — surface every tag (license:apache-2.0, license:mit, license:llama3, …) across a curated list

How to use

1. Enter a Search query (e.g. llama, stable-diffusion, whisper) — or leave empty to browse the whole Hub
2. Optionally pick a Pipeline task to narrow to a single task type
3. Optionally enter a Library filter (transformers, diffusers, sentence-transformers, peft, timm, …)
4. Pick a Sort by field (Downloads, Likes, Recently updated, Recently created, Trending) and a Sort direction
5. Set Max items — the actor auto-paginates the Hugging Face Hub API to collect them
6. Run the actor — one model per row in the Dataset tab

Output format

One model per dataset row — perfect for direct CSV, Excel, or Google Sheets export:

{
  "id": "meta-llama/Llama-3.1-8B",
  "author": "meta-llama",
  "modelName": "Llama-3.1-8B",
  "huggingfaceUrl": "https://huggingface.co/meta-llama/Llama-3.1-8B",
  "pipelineTag": "text-generation",
  "libraryName": "transformers",
  "tags": ["transformers", "safetensors", "llama", "text-generation", "en", "license:llama3.1"],
  "downloads": 2845123,
  "likes": 5421,
  "gated": "manual",
  "private": false,
  "createdAt": "2024-07-22T18:30:00.000Z",
  "lastModified": "2025-09-14T10:12:33.000Z",
  "trendingScore": 87.4,
  "scrapedAt": "2026-05-23T12:34:56.000Z"
}

Input options

Field	Type	Description
Search query	string	Free-text search across model IDs and descriptions — optional
Pipeline task	enum	Restrict to a Hugging Face pipeline tag (text-generation, text-to-image, …) — optional
Library filter	string	Restrict to a single library (transformers, diffusers, sentence-transformers, …) — optional
Sort by	enum	Downloads, Likes, Recently updated, Recently created, Trending — default Downloads
Sort direction	enum	Descending (most first) or Ascending (least first) — default Descending
Max items	integer	Total models to collect across paginated calls — default 100, max 5000