arXiv
Follow categories and authors, get summarized abstracts.
arXiv publishes hundreds of preprints a day. Pile follows the categories and authors you choose, fetches every new paper and summarizes the abstract in plain language. You spot the paper that matters without reading fifty cryptic titles.
Extracted, ready to revisit
- The title and authors of the preprint, with a direct link to the arXiv page.
- The full abstract, kept as-is alongside the plain-language summary.
- The main contribution identified by your AI: what the paper concretely brings.
- The arXiv category (cs.LG, cs.CL, stat.ML...) and keywords, so you can filter by subfield.
- The submission date and, when available, revised versions, so you can track a paper's evolution.
Connect arXiv
- 1
Pick your categories
In Connect, pick arXiv and list the categories to follow (for example cs.LG, cs.CL) or specific authors.
- 2
Pile watches new papers
On each new preprint in your scope, Pile fetches title, authors and abstract.
- 3
Summarized abstract
Your AI rewrites the abstract in plain language, flags the main contribution and tags by subfield.
- 4
Wiring being finalized
The connector is coded and tested. Runtime wiring (keys, scheduling, ingest queue) is being finalized. Sign up and we open this source for you the moment it ships, no second signup.
Concretely, what for
Watch cs.LG without drowning in daily submissions
cs.LG sometimes lists over a hundred new papers a day. Pile fetches each title and abstract, summarizes them in one or two sentences and scores them. In five minutes you know which papers deserve a full read today.
Follow a researcher or a research group
You care about the work of a specific team. Pile watches their new arXiv submissions by author name and notifies you on each new paper, summarized. No more manually checking the arXiv site.
Build a living research bibliography
You are working on a topic for several weeks. Pile accumulates relevant papers in a dedicated vault, summarized and tagged. Your bibliography grows with each new submission without any manual action on your part.
Practical tips
- Combine broad categories and specific authors: you catch the whole field plus the researchers you follow.
- The plain-language summary saves you from decoding twenty dense abstracts to find the right paper.
- File a research topic into a vault: you build a living bibliography.
If something breaks
I am not receiving new papers for my category
Check that the category is entered exactly as arXiv names it (examples: cs.LG, cs.CL, math.ST). Categories are case-sensitive. If the problem persists, remove the category from your list and re-add it in Connect.
An author I follow is not being recognized
Pile searches by author name as it appears in arXiv submissions. Try variants (full first name, initials, name with or without diacritics). If an author publishes under several forms of their name, add each variant separately.
The summarized abstract seems imprecise on a highly technical paper
Heavily formal papers with dense mathematical notation are harder to summarize. Pile keeps the original abstract alongside the summary: check it to verify accuracy, and open the full PDF if the topic is critical for you.
Frequently asked
Can I follow a specific author?
Yes. You can follow categories, authors, or both at once.
Does Pile download the full PDF?
Pile summarizes the abstract and keeps the link to the PDF. You open the full paper only if the summary hooks you.
Available when?
The connector is coded; runtime wiring is being finalized. Sign up for the opening.
Keep filling your Pile
Ready to connect arXiv?
Connect your first source in 30 seconds. No card, no account to try.