arXiv provider

The arXiv provider exposes papers as objects. A paper directory contains JSON metadata plus paper and source artifacts. Category paths list papers by arXiv category.

No credentials are required.

At a glance

Field	Value
Provider id	arxiv
Default mount	/arxiv
WASM	omnifs_provider_arxiv.wasm
Manifest	providers/arxiv/omnifs.provider.json
Routes	providers/arxiv/src/lib.rs
Auth	none
Capabilities	export.arxiv.org and arxiv.org for metadata and paper resources; declared memory need 64 MB

Setup

omnifs init arxiv
omnifs up
omnifs shell

Main paths

/arxiv/papers/{id}
/arxiv/papers/{id}/paper.json
/arxiv/papers/{id}/paper.pdf
/arxiv/papers/{id}/source.tar.gz
/arxiv/papers/{id}/versions
/arxiv/papers/{id}/versions/v{n}
/arxiv/papers/{id}/versions/v{n}/paper.json
/arxiv/papers/{id}/versions/v{n}/paper.pdf
/arxiv/papers/{id}/versions/v{n}/source.tar.gz
/arxiv/categories/{category}
/arxiv/categories/{category}/papers

The provider also attaches the same paper subtree under category paper listings, so a paper reached from a category has the same readable leaves as one reached from /papers/{id}.

Examples

ls /omnifs/arxiv/papers/1706.03762
cat /omnifs/arxiv/papers/1706.03762/paper.json | jq .title
cp /omnifs/arxiv/papers/1706.03762/paper.pdf .
ls /omnifs/arxiv/categories/cs.LG/papers
cat /omnifs/arxiv/papers/1706.03762/versions/v1/paper.json | jq .

Capability and cache notes

The provider declares arXiv domains and expected memory needs. Metadata reads can populate canonical paper objects and rendered leaves. Large paper/source bytes are served through host-managed byte sources rather than as arbitrary provider memory.

Limitations

Category paths list recent papers for a category. They are not a general arXiv search interface. Versioned artifacts use /versions/vN; old-style arXiv ids with / must be encoded into one path segment.