HuggingFace Paper Publisher logo

HuggingFace Paper Publisher

Visit

Hugging Face's paper-publisher skill indexes arXiv papers on the Hub, links them to models and datasets, claims authorship and drafts articles.

Share:
View alternatives

HuggingFace Paper Publisher is the huggingface-paper-publisher skill in the official huggingface/skills repository (skill folder). It gives a coding agent a script, paper_manager.py, plus templates for the research side of the Hub: getting a paper onto Hugging Face Paper Pages, connecting it to the models and datasets it describes, and managing authorship.

The script declares its dependencies inline (PEP 723), so the skill tells agents to run it with uv run instead of building an environment.

Key Features

  • Index a paper: uv run scripts/paper_manager.py index --arxiv-id "2301.12345", or check to see whether it already exists.
  • Link to artifacts: link --repo-id user/model --repo-type model --arxiv-id ... adds the citation to a model or dataset README. Several IDs can be passed with --arxiv-ids.
  • Automatic tagging: once a README links an arXiv paper, the Hub adds an arxiv:<id> tag and the Paper Page lists every repo that cites it.
  • Authorship: claim, check-authorship, list-my-papers and toggle-visibility manage which verified papers show on your profile. Claims are verified by Hugging Face staff.
  • Article drafts: create --template standard|modern|arxiv|ml-report generates a markdown research article with a table of contents and LaTeX math.

Use Cases

  • Releasing a model and its paper together so each page links to the other.
  • A lab adding its back catalog of arXiv papers to the Hub.
  • Writing an experiment report in a consistent template.

Pricing

Free and open source (Apache-2.0 repository). Paper Pages and repo links need only a Hugging Face account and a write token.

Getting Started

  1. Install HuggingFace CLI, then hf skills add huggingface-paper-publisher.
  2. Export HF_TOKEN with write access.
  3. Ask the agent: "index arXiv 2301.12345 and link it to my model".
  4. Review the README diff before it lands on the Hub.

Limitation: the skill edits README metadata on real repos, so run it against the right --repo-id and review changes. The article templates format a paper; they do not check its content, citations or claims.

FAQ

Can it upload a PDF that is not on arXiv?

The documented flows start from an arXiv ID.

Does claiming authorship happen instantly?

No. The skill notes that the Hugging Face team verifies claims.

Alternatives

Conclusion

Useful when your research outputs live on the Hub and you want papers, models and datasets cross-linked with little manual editing. More in the skills hub.

Comments

No comments yet. Be the first to comment!