Think inside your AI world.
How do you actually keep track of new papers in your field?
Keeping up with the literature is the task everyone means to do and nobody has a good habit for. The feeds are clean and free, arXiv and Crossref both keyless. The problem was never access; it is that a keyword alert floods you with matches while the two or three papers that actually bear on your work sit in the pile unread.
The construction is the open paper feeds plus your own relevance rule. arXiv and Crossref are keyless; between them they carry new preprints, their categories, abstracts and citations. Your rule is “my categories and my keywords, or anything that cites the three papers my work is built on.” Through Unl the handful that meet that rule arrive beside it when your own research comes up, stamped with when they were posted. The feed was always open; the filter that fits your work is the part that was missing.
The person
A researcher or practitioner tracking a subfield who wants to stay current without turning it into a daily chore. They know which categories are theirs, which keywords define their corner, and which foundational papers their own work stands on.
The pieces, and how far each one reaches
| Source | What it gives you | Reach today |
|---|---|---|
| arXiv API | New preprints with their categories, abstracts and authors, so a paper can be matched to your categories and your keyword set. | Keyless and clean, with clear category taxonomies; one of the tidiest public feeds in this whole set. |
| Crossref API | Published works and their citation metadata, so a new paper citing the work you build on can be caught even outside your usual categories. | Keyless. Coverage is broad but not total, and metadata quality varies by publisher, which is graded honestly rather than assumed. |
The honest limits are coverage and metadata, not access. arXiv is clean and keyless; Crossref is keyless and broad but not exhaustive, and citation metadata quality varies by publisher, so a citation rule catches most of what cites your foundations, not provably all of it. Both feeds are verified at paste. The value is a measured read against your own relevance rule, not a claim to have indexed every paper ever written.
The rule, in their words
- “New in my arXiv categories, and matching my keyword set.”
- “Anything that cites the three papers my work is built on.”
- “From the ten authors whose work I always read.”
The compound is what a keyword alert cannot do: category AND keyword, OR a citation to your foundations, OR an author you follow. A single keyword alert either floods you or misses the paper that never uses your keyword but cites the work you stand on. The rule that braids topic, citation and author is the one that surfaces the handful worth your afternoon.
Unprompted
Your AI
Since it is on your mind, three crossed your rule this week: your categories and keywords, or a citation to your three foundations. Two new preprints in your arXiv categories match your keyword set, and a third, outside your usual categories, cites the paper your own method is built on.
That is from the arXiv and Crossref feeds, both keyless, stamped with when each was posted. The citation one is the sort a plain keyword alert would have missed.
The papers arrived against the researcher’s own braided rule, topic and citation and author, not a raw keyword dump. Which to actually read is theirs.
Why this beats the dashboard
A keyword alert emails you every match and treats a passing mention like a central result; it cannot follow a citation to the work you stand on. Your rule braids topic, citation and author, and it surfaces when your own research is the subject rather than as a daily digest you learn to archive unread. The handful that reach you are already measured against what genuinely bears on your work.
Keeping track of new papers in your field is the open paper feeds sifted by your own relevance rule; through Unl the handful that meet “my categories and keywords, or a citation to my foundations” arrive beside that rule, stamped with when they were posted, and your relevance rule does the choosing, not a raw keyword match.
Reads through Unl arrive with measured context — in the presence of the decisions you’ve already settled. The reach lane is live: one box, paste anything. If it speaks MCP, Unl can reach it. Readings arrive unprompted, the data beside the criterion; Unl is a courier, not a warehouse, and keeps only your keys and the frame.
Read further
Questions people ask
Which paper feeds does this use, and are they free?
arXiv and Crossref, both keyless. arXiv is one of the cleanest public feeds anywhere, with clear category taxonomies; Crossref is keyless and broad but not exhaustive, and its citation metadata quality varies by publisher. Access was never the problem for keeping up with the literature; the missing piece was a filter that fits your own work, which is what your rule provides.
How is this better than a keyword alert?
A keyword alert floods you with every match and cannot follow a citation. Your rule is compound, your categories and keywords, or a paper that cites the three your work is built on, or one of the authors you always read, and it surfaces the two or three that actually bear on your work rather than a daily digest. The braid of topic, citation and author is the part no single keyword can express.
Does Unl summarise or rank the papers for me?
Unl brings the papers that meet your rule beside it, stamped with when they were posted, and stops. It does not decide which is most important or tell you what to read first; the relevance rule is yours and the reading of the papers is yours. Bringing the right handful into view is the whole job.
What this is
Think inside your AI world — you stay in command
Unlimitless (Unl to friends) holds what you've settled, reads what your tools are showing, and catches what's changed out in the world — and hands your AI whatever bears on the work, the moment it's needed, without you asking. The right thing, in front of the model, unprompted, with you in command of the call. So you keep moving toward what you set out to build, on top of everything you've already decided.
It plugs into Claude, Claude Code, ChatGPT and Cursor as an MCP connector. Quick to connect, in a couple of steps.
Unlimitless is open now to invited Alpha. Apply for the Beta waitlist to come in ahead of the full launch:
Alpha is invite-only · free at launch.