About unkwn42.

I'm Rohit Pawar. During my data science degree my supervisor gave me a dissertation topic that, it turned out, had more or less been done before. I spent months finding the papers and not much time on the research itself.

Other researchers have told me versions of the same story, and theirs are worse: they found out from a referee, or a year in. I'd like people to find out before they submit.

What it does

A plagiarism checker makes sure you didn't copy anyone. unkwn42 makes sure you didn't miss anyone. You paste your abstract and reference list, I add the papers that cite the same sources you do, and the check searches those plus the 438,284 climate and earth science papers we hold from OpenAlex. You get three answers:

• Is my main claim already published? Claimed gives you the paper that already makes it. Adjacent says what the nearest work did and where yours differs. Open means we didn't find it, which tells you about our corpus, not proof that nobody has done it.
• What am I missing? Papers a referee will expect you to cite that aren't in your list, each with why it matters and how yours differs.
• Is my reference list clean? Retracted papers, corrections, preprints that have since been published, and years or DOIs that don't match.

Every paper it shows is real and links to its source. The first check is free; the full check is $19 a paper, with 30 days of re-checks. A Claimed result isn't a rejection, by the way. Replicating or extending known work is real research. It just helps to know where you stand before a referee tells you.

Why other words matter

Every field names things its own way. In the papers we hold, nobody writes "marine heatwave" before 2016, and by 2019 it's the main term, though the events are much older. You can't search for a word you don't know, so the check looks for your claim under other names too. Renamed: before they were marine heatwaves has the counts.

What it gets wrong

It will miss things. In the eleven areas the corpus was built from we hold between about a third and three fifths of the last ten years of papers, and much less in hydrology, wildfire or marine biology. The papers added for each check close part of that gap, and every result says how much of your reference list we hold. When there isn't enough to judge, it says so instead of guessing.

If a paper should have shown up, there are buttons on every result to tell me. I trace each miss back to where the search lost it, and it becomes a test the next version has to pass.

Who it's for

PhD students and postdocs in climate and earth science who are about to submit a paper (more here). Later I'd like journals to run it on every submission; for journals says how that would start.

The company

unkwn42 is run by UNKWN42 LTD, registered in England and Wales (number 17207152). It's in beta and I'm building it on my own, so if you write in, you'll hear back from me.

Email hello@unkwn42.com or use the contact form.