Research writing is formal, structured and citation-dense by design. ScribeLens is calibrated for exactly that: it separates the parts of a paper that carry your voice from the apparatus that surrounds them, and shows you the evidence behind every result.
A bibliography is the most formulaic text in any paper. Rigid citation formatting repeats the same structures hundreds of times, which is a poor basis for judging authorship.
ScribeLens detects and carves off reference lists before scoring, along with in-text citation apparatus, DOI and ISSN lines, journal running headers, and page-number furniture. The excluded content stays visible in your report for transparency, but never influences the verdict. No manual preparation is needed: paste or upload the whole paper.
ScribeLens analyses up to 40,000 words in one pass, which covers a full research paper or a thesis chapter without splitting it into sections.
That matters because context changes results. A detector that only sees a couple of thousand words at a time judges your introduction without ever seeing your methods, and you are left stitching together fragments of an answer.
Every result breaks down to the sentence: which passages carried signals, which reasons were triggered, and how the overall interpretation follows from them.
You can generate a verification link so a supervisor, co-author or editor can open the exact report you saw, unedited, at its own URL. Reports also export to PDF and DOCX with the analysed text and highlights intact.
Formal register, hedged claims, passive constructions and standardised section headings are the conventions of the field, not signals of authorship. ScribeLens is tuned on published academic writing so that scholarly style is read as scholarly style.
Where a result is genuinely uncertain, the report says so plainly and shows the share of sentences behind it, rather than resolving ambiguity into a single confident number.
Literature reviews are citation-dense and formal, which is exactly the profile ScribeLens is calibrated for. Citation apparatus is excluded before scoring, and the report shows you sentence by sentence what contributed to the result, so you can see precisely what any signal is based on.
No. Reference lists, in-text citation apparatus and declaration sections are detected and excluded automatically. Paste or upload the whole paper, including the bibliography.
Yes. Create a verification link and the stored report opens at its own URL, exactly as you saw it. You can also export to PDF or DOCX, and every report carries a scan ID that can be checked against the stored copy.
It means the sentence-level evidence points in more than one direction, and the report says so rather than forcing a verdict. It shows the share of sentences reading clearly human, sitting in between, and carrying strong AI signals, so you can review the specific passages involved and judge them in context.
Up to 40,000 words in one pass, with references excluded automatically and sentence-level evidence for every result.