Explore Science vs Penelope.ai

Penelope takes a Word file, reads it against a specific journal's author guidelines, and returns an annotated report: is there a dedicated ethical approval section with the title the journal wants, is a funder named, is the abstract structured with the subheadings PRISMA or CONSORT expects, are the figures numbered in ascending order, is the word count inside the limit, are there track changes or Endnote field codes still in the file. Every check links to the passage it fired on, and the pass and fail criteria are public. As a way of not losing three weeks to an editorial office bounce, it works, and on a journal page that carries it the report costs the author nothing.

What the list does not contain is a check on the research. No check reads the design, asks whether the sample supports the inference, or notices that the discussion claims more than the results earned. Explore Science runs at that level: methods, results and argument read end to end by several models that then check each other, and a Calibre score out of 100 across eight fixed categories, so the second draft can be measured against the first on the same axes rather than described as better.

Penelope's referencing checks are internal ones. Does every citation in the text appear in the reference list, does every reference get cited, is the style numerical or named as the journal requires, is the ordering right. Those are the errors a copy editor would otherwise chase, and catching them before submission is worth something.

None of them leaves the document. Explore Science resolves each reference against live scholarly databases and reports back whether the work exists, whether it is attributed correctly, and whether it says what the sentence citing it needs it to say. A reference list can be perfectly consistent and still rest on a paper that found the opposite result.

The product is sold to journals: £750 a year under 500 submissions, down to £0.80 a submission at scale, with the editor choosing which checks run, which are critical, and how the feedback is worded. Its natural home is the submit button, and journal-branded pages exist for titles including Access Microbiology and SPIE's Journal of Electronic Imaging. If your journal is on Manuscript Manager you may already be using it without having chosen it; Editorial Manager and ScholarOne are on a waiting list.

That placement is its strength and its ceiling. A check tuned to one journal's guidelines is useful at the moment you submit to that journal and inert before then, when the paper is still being argued into shape and the target may change. Explore Science works on the draft that has not been assigned a venue yet, and on the protocol before the study runs.