Hugging Face scraper API
The InsightSocial Hugging Face API has 42 endpoints, from 20 credits a call. A model call costs 20 credits, about $0.02 on Pro, and your first 10 calls up to 100 credits each are free.
The Hugging Face Hub as JSON: models with downloads, cards and base models, datasets and their rows, Spaces, papers and daily papers, collections, accounts and repository discussions.
42 endpoints · from 20 credits a call · 10 free calls up to 100 credits each · no credit card
curl "https://api.insightsocial.app/v1/huggingface/model?owner=openai-community&name=gpt2" \-H "x-api-key: $INSIGHTSOCIAL_API_KEY"
What Hugging Face data can you get?
Public Hugging Face data only, in the same JSON envelope as every other platform. Every Hugging Face endpoint is beta: inside data it returns Hugging Face's own structure under our naming rules (snake_case keys, string ids, ISO times where Hugging Face gives a timestamp), not the shared profile and post schema, and fields may change.
Models
A model with downloads, likes, tags, card, weights, base models and evaluation results, and model search by task, library, licence, base model and size.
Datasets
A dataset with its card and downloads, its rows a page at a time, per-column statistics, and dataset search by language, licence and tags.
Spaces, papers and collections
Spaces and their hardware, search by plain-words description, papers with full text and the repos citing them, the daily papers list, and curated collections.
Accounts, repos and trending
An account and its repos, followers and activity, a repo's commits, files, README, likers and discussions, and what is trending now.
What people build with the Hugging Face API
Model landscape research
Search models by task and licence, sort by downloads, and read each one's base model and derived fine-tunes to map a family.
Paper tracking
Pull the daily papers every morning, read the full text of the ones that matter, and list the models and datasets that cite them.
Dataset sampling
Read a dataset's card, column statistics and a few pages of rows before you download it.
Hugging Face API endpoints
All 42 Hugging Face endpoints and their price in credits. Every one has a fixed price.
| Endpoint | Credits | What it returns |
|---|---|---|
| /account | 20 | Get a Hugging Face user or organization with its counts, organizations and profile text. |
| /account/activity | 20 | List the recent public activity of a Hugging Face user or organization. |
| /account/buckets | 20 | List the public storage buckets of a Hugging Face account. |
| /account/followers | 20 | List the followers of a Hugging Face user or organization. |
| /account/profile | 20 | Get Hugging Face profile details beyond the account record. |
| /account/repos | 20 | List the public models, datasets or Spaces of a Hugging Face account. |
| /blog/posts | 20 | List posts from the official or community Hugging Face blog. |
| /changes | 20 | List the Hugging Face models, datasets, Spaces, papers or blog posts that changed most recently. |
| /collection | 20 | Get a Hugging Face collection with every item and the curator's notes. |
| /collections | 20 | Search Hugging Face collections, or find the collections that contain a model, dataset, Space or paper. |
| /dataset | 20 | Get a Hugging Face dataset with its downloads, tags, description and card. |
| /dataset/rows | 20 | List rows from a Hugging Face dataset a page at a time. |
| /dataset/statistics | 20 | Get per-column statistics of a Hugging Face dataset split. |
| /dataset/viewer | 20 | Get what a Hugging Face dataset viewer can show. |
| /datasets | 20 | Search and filter Hugging Face datasets by keyword, language, licence and tags. |
| /inference-providers/models | 20 | List the Hugging Face models that inference providers host. |
| /kernel | 20 | Get a Hugging Face compute kernel. |
| /kernels | 20 | Search Hugging Face compute kernels. |
| /model | 20 | Get a Hugging Face model with its downloads, tags, card, weights, base models and evaluation results. |
| /models | 20 | Search and filter Hugging Face models by task, library, language, licence, base model and more. |
| /organization/members | 20 | List every member of a Hugging Face organization. |
| /paper | 20 | Get a Hugging Face paper with its authors, abstract and upvotes. |
| /paper/content | 20 | Get the full text of a Hugging Face paper as Markdown. |
| /paper/repos | 20 | List the Hugging Face models, datasets and Spaces that cite a paper. |
| /papers | 20 | Search Hugging Face papers by keyword, or list the newest papers. |
| /papers/daily | 20 | List the papers on the Hugging Face daily papers list for a day, week or month. |
| /posts | 20 | List posts from the Hugging Face social feed, newest first. |
| /repo/commits | 20 | List the commit history of a Hugging Face repository. |
| /repo/discussion | 20 | Get one discussion or pull request of a Hugging Face repository with every comment. |
| /repo/discussions | 20 | List the discussions and pull requests of a Hugging Face repository. |
| /repo/files | 20 | List the files and folders of a Hugging Face repository. |
| /repo/likers | 20 | List the accounts that liked a Hugging Face repository. |
| /repo/readme | 20 | Get the README of a Hugging Face model, dataset or Space. |
| /search | 20 | Search Hugging Face models, datasets, Spaces, organizations, users, papers, collections, buckets, kernels and blog posts. |
| /search/readmes | 20 | Search the text of Hugging Face READMEs and code files. |
| /space | 20 | Get a Hugging Face Space with its run state, hardware, SDK and linked models and datasets. |
| /spaces | 20 | Search Hugging Face Spaces by keyword, SDK and tags. |
| /spaces/semantic-search | 20 | Search Hugging Face Spaces that match a plain-words description. |
| /tags | 20 | List the Hugging Face tags in one group, such as tasks or languages. |
| /trending | 20 | List the models, datasets and Spaces trending on Hugging Face now. |
| /user/following | 20 | List the accounts a Hugging Face user follows. |
| /user/likes | 20 | List the repositories a Hugging Face user liked. |
Parameters, response fields and examples for each endpoint are in the Hugging Face API reference.
InsightSocial vs the official Hugging Face API
The Hugging Face Hub API is free and documented, with Python and JavaScript clients; public repos can be read without a token, and a token lifts limits and opens gated repos you have accepted.
| InsightSocial API | Official Hugging Face API | |
|---|---|---|
| Setup | Sign in and copy a key | Optional Hugging Face token |
| Gated repos | Public metadata only | With a token, after you accept the terms |
| Other platforms | The same key and envelope for 60 more | Hugging Face only |
| Cost per call | 20 credits | Free, within Hugging Face's limits |
The official API is the right tool for posting, messaging or managing accounts you own. This one is read-only public data.
How much does the Hugging Face API cost?
Hugging Face calls start at 20 credits. Every endpoint's price is in the table above and in the free GET /v1/endpoints catalogue, so you know the cost before you call. Pro's 10,000 monthly credits cover about 500 model lookups at 20 credits each.
Failed calls are free
A call that errors or finds nothing costs 0 credits.
Repeats and retries
A repeat call is charged again: the full price when it fetches new data, 2 credits when the answer comes from the shared cache. Add fresh=1 to force a new read. Only a retry with the same Idempotency-Key is free.
One balance
API calls and Chrome extension exports share the same credits. The free plan includes 500 a month.
Not writing code?
Connect InsightSocial to Claude or ChatGPT and ask for Hugging Face data in plain English. Same account, same credits.
Hugging Face API: frequently asked questions
If Hugging Face is your only source, the Hub API and its clients are free and the natural choice. This API suits a pipeline that also reads GitHub, Hacker News, Reddit or X: one key, one envelope and one credit balance across all of them.
/v1/huggingface/model takes the owner and the name separately, for example owner=openai-community and name=gpt2. A bare id without its owner is not accepted.
/v1/huggingface/dataset/rows returns rows a page at a time for a config and split, and /v1/huggingface/dataset/statistics returns per-column statistics.
/v1/huggingface/papers/daily returns the daily papers list for a day, week or month, and /v1/huggingface/paper/content returns a paper's full text as Markdown.
They return Hugging Face's own record shapes under our naming rules (snake_case keys, string ids, ISO times) rather than the shared profile and post entities, so fields may change. Every Hugging Face endpoint is 20 credits a call. Calls that fail or find nothing are free.
No. There is no Hugging Face app, OAuth flow or review. One API key in the x-api-key header works for Hugging Face and the other 60 platforms.
Yes. Your first 10 calls priced at 100 credits or less are free, and the free plan includes 500 credits every month. No credit card is needed to get a key.
60 requests a minute and 10 requests in flight at once, per API key. Going over either returns a 429 with a Retry-After header, and a rejected request costs nothing. There is no daily quota: past those two limits, your credit balance is what bounds volume.
List endpoints return rows in data.items and a pagination object with next_cursor and has_more. Send next_cursor back unchanged as the cursor parameter for the next page, and stop when has_more is false. A handful of endpoints page with their own parameter instead; their reference pages name it.
No. A call that errors or finds nothing costs 0 credits. A live fetch costs the endpoint's price, or for a metered endpoint what it actually read. A repeat call is charged again: the full price when it fetches new data, 2 credits when the answer comes from the shared cache. Add fresh=1 to force a new read. Only a retry with the same Idempotency-Key is free.
A call fetches live unless a recent copy is in the shared cache, and every response says which with the cached field; a cached answer costs 2 credits. Add fresh=1 or send Cache-Control: no-cache to skip the cache; that call is charged at the normal price.
Collecting publicly available data is generally treated differently from accessing private data, but the answer depends on your jurisdiction, Hugging Face's terms and what you do with personal data afterwards. The API reads public data only. Talk to a lawyer for anything commercial.
Hugging Face API guides
The same key works for 60 more platforms
- TikTok scraper API
- LinkedIn scraper API
- Twitter/X scraper API
- YouTube scraper API
- Reddit scraper API
- Pinterest scraper API
- Bluesky scraper API
- Truth Social scraper API
- Snapchat scraper API
- Telegram scraper API
- Twitch scraper API
- Kick scraper API
- Rumble scraper API
- Kwai scraper API
- Douyin scraper API
- Xiaohongshu (RedNote) scraper API
- Quora scraper API
- TikTok Shop scraper API
- Weibo scraper API
- Zhihu scraper API
- Substack scraper API
- Nextdoor scraper API
- Amazon scraper API
- App Store scraper API
- Apple Music scraper API
- Behance scraper API
- Dailymotion scraper API
- Deezer scraper API
- DEV Community scraper API
- Discord scraper API
- Dribbble scraper API
- Fandom scraper API
- Flickr scraper API
- GitHub scraper API
- Google Ads Transparency scraper API
- Google News scraper API
- Google Trends scraper API
- Gumroad scraper API
- Hacker News scraper API
- Ko-fi scraper API
- Lemmy scraper API
- Letterboxd scraper API
- Linktree scraper API
- Lnk.Bio scraper API
- Mastodon scraper API
- Medium scraper API
- MyAnimeList scraper API
- npm scraper API
- Polymarket scraper API
- Product Hunt scraper API
- Shopify scraper API
- SoundCloud scraper API
- Spotify scraper API
- Stack Overflow scraper API
- Stan Store scraper API
- Steam scraper API
- Vimeo scraper API
- Wikipedia scraper API
- Yahoo Finance scraper API
- Yelp scraper API