Hugging Face Papers API Documentation

This document describes the public API endpoints that power the Daily Papers feature on huggingface.co. These endpoints provide access to research papers, metadata, and daily-curated listings derived from arXiv and other scientific sources.

The API is publicly available for read-only operations and is used internally by the Hugging Face frontend (built with SvelteKit) to render the Papers interface.

Checkout the app wrapped around this API HF Papers App

---

Overview

The Daily Papers API enables:

• Keyword-based paper search
• Retrieval of detailed paper metadata
• Paginated access to daily-curated papers

All endpoints return data in JSON format and are accessible without authentication.

---

Base URL

https://huggingface.co

All API routes are relative to this base domain.

---

Available Endpoints

Endpoint	Method	Description	Pagination
/api/papers/search	GET	Search for papers by keyword.	No
/api/papers/{arxiv_id}	GET	Retrieve details for a specific paper.	N/A
/api/daily_papers	GET	Retrieve curated papers for a given date.	Yes (page, limit)

---

1. GET /api/papers/search

Description

Searches the papers database by title, abstract, or author. This endpoint returns a limited number of results (usually 20–25). Pagination parameters are currently not supported.

Example Request

GET https://huggingface.co/api/papers/search?q=transformer

Query Parameters

Parameter	Type	Required	Description
q	string	Yes	Search keyword (title, abstract, or author).
limit	integer	No	Currently ignored.
offset	integer	No	Currently ignored.

---

2. GET /api/papers/{arxiv_id}

Description

Retrieves metadata and detailed information for a specific paper by its arXiv ID.

Example Request

GET https://huggingface.co/api/papers/2411.04567

Path Parameters

Parameter	Type	Required	Description
arxiv_id	string	Yes	The arXiv identifier for the paper.

---

3. GET /api/daily_papers

Description

Retrieves a paginated list of curated papers for a specific date. This endpoint supports pagination through the page and limit parameters.

Example Request

GET https://huggingface.co/api/daily_papers?date=2025-11-10&page=1&limit=1

Query Parameters

Parameter	Type	Required	Description	Example
date	string	Yes	Date for which to fetch papers (YYYY-MM-DD).	2025-11-10
page	integer	No	Page number, starting from 1.	1
limit	integer	No	Number of results per page (default: 10–20).	5

Pagination

The /api/daily_papers endpoint implements standard pagination.

Example Integration

JavaScript Example

async function getDailyPapers(date, page = 1, limit = 5) {
  const response = await fetch(
    `https://huggingface.co/api/daily_papers?date=${date}&page=${page}&limit=${limit}`
  );
  const data = await response.json();
  return data.results;
}

async function getPaperDetails(arxivId) {
  const response = await fetch(`https://huggingface.co/api/papers/${arxivId}`);
  return await response.json();
}

Python Example

import requests

def get_daily_papers(date, page=1, limit=5):
    url = f"https://huggingface.co/api/daily_papers?date={date}&page={page}&limit={limit}"
    response = requests.get(url)
    return response.json()

def get_paper_details(arxiv_id):
    url = f"https://huggingface.co/api/papers/{arxiv_id}"
    response = requests.get(url)
    return response.json()

---

Notes

• The API is public and does not require authentication.
• Responses are cached by the Hugging Face CDN for performance.
• Data originates primarily from the arXiv API.
• /api/papers/search is optimized for interactive client-side use and may not expose pagination.

---

License

Copyright licence by HF is added here

---

Authors

This documentation was prepared based on public behavior of the Hugging Face Papers API and the open endpoints available under huggingface.co. Authored and maintained by contributors analyzing the Hugging Face frontend and public Hub API.