AI Research Firehose
Recent arXiv AI/ML, raw, fast, unverified. One static file. No login, no paywall.
Machine-readable JSON/api/papers/arxiv-recentRead this first. A raw, unranked, unfiltered stream of the most recent arXiv AI/ML submissions. Every field is copied verbatim from the arXiv API. There is no editorial review, no dedup against a curated index, and no quality gate, by design. If your agent needs a citation-ranked view of the field, use /api/papers/ai-trending; for an editor-curated daily set, use /api/papers/hf-daily.
What this is
A flat dump of recent arXiv AI/ML papers (cs.AI, cs.LG, cs.CL, cs.CV, cs.RO, cs.NE, cs.IR, stat.ML). Each record carries the paper's bibliographic facts copied verbatim from the arXiv API. There is no quality gate on this feed by design. It is not ranked, not deduped against a curated index, and not editorially reviewed.
Fields in each record
Verbatim from arXiv
- arxivId
- version
- title
- abstract
- authors
- primaryCategory
- categories
- publishedAt
- updatedAt
- htmlUrl
- pdfUrl
- doi
Author affiliations are almost entirely "unknown" (an arXiv API limitation); treat them as absent.
The feed
GET /api/papers/arxiv-recentfreeOne JSON snapshot, refreshed daily at 11:30 UTC from the arXiv API. The response is { ok, snapshot }, and snapshot.papers is the array of records. There are no query parameters; it is not filterable server side, on purpose.
curl 'https://tensorfeed.ai/api/papers/arxiv-recent'
FAQ
Is this a TensorFeed quality feed?
No. It is a raw firehose. Every field (arXiv id, version, title, abstract, authors, primary category, all categories, published and updated dates, HTML and PDF links, doi) is copied verbatim from the arXiv API and is sound. There is no interpretive layer: no subfield, methodology, or milestone tagging. It is the unranked stream of what just landed.
Is it ranked, deduped, or editorially reviewed?
No, by design. The feed is sorted by submission date only. It is not ranked by importance, not deduped against a curated index, and not editorially reviewed. For a citation-ranked view of the field, use /api/papers/ai-trending; for an editor-curated daily set, use /api/papers/hf-daily.
Can I filter the feed server side?
No. The endpoint returns the latest snapshot whole, with no query parameters. Read it and filter client side. It refreshes once per day from the arXiv API.
Is there a paid tier?
No. This data is free only. There is no premium tier on the research firehose, ever.