👋 Product Hunt 25% off lifetime with code PHUNT25 ·
this week only
Students now generate work with AI – complete with fabricated, misattributed, or overstated citations. Acurio checks the citation fidelity of whole batches of theses: does every claim actually match the source it cites? Automatically, traceably, in minutes instead of days.
Recent meta-analyses show a marked reduction in cognitive load under guided feedback (Müller & Chen, 2021). The effect is most stable in cohorts above n=120, though long-term effects across language groups remain inconsistent (Vega, 2019). Several authors claim a complete elimination of the test plateau in adults (Hartmann, 2018) — a finding our re-analysis cannot reproduce.
“No comparable statement found in Hartmann 2018. The closest passage cites a 4% improvement (p. 88) — no plateau elimination.”
This is how an Acurio report reads: every citation color-coded, with a verdict and the exact source passage — right in the document.
Plagiarism tools find copied text. They do not find whether a source even exists, whether it actually supports the claim, or whether it was attributed correctly. Acurio asks a different question.
Was this text copied from somewhere?
Finds text similarity.
Compares against the web.
Does the cited source really say what's claimed here?
Finds fabricated, misattributed, and overstated citations.
Compares each claim against the uploaded source.
Acurio doesn't replace your plagiarism check — it closes the gap that one leaves open.
Checking the citation fidelity of a single thesis by hand easily costs several hours. At 30, 100, or 400 theses a semester it isn't feasible — which is precisely why it doesn't get done.
Drop in theses (.docx / .tex) and their sources (PDF / RIS / BibTeX), in batches.
Acurio extracts every citation and the claim behind it, then locates the relevant passage in the source.
Each claim gets a traceable verdict: supported, partial, unsupported — with the exact passage.
One clear report per thesis, flagged passages first. Examiners read minutes instead of hours.
No workflow overhaul. No training. If you can upload a file, you can run Acurio.
The report surfaces the few problematic citations — you don't re-read the whole thesis.
Every verdict carries the exact source passage. Defensible to students and examination boards.
Where a source is missing, Acurio says so clearly — instead of falsely marking a claim unsupported.
Flagged passages can be reviewed in place and the thesis marked as checked.
Every thesis is checked to the same standard, regardless of who supervised it.
Licenses per chair, institute, or faculty; centrally managed.
Batches from dozens to hundreds of theses during submission waves — without extra headcount.
EU hosting (Germany), GDPR-compliant, your content is never used for model training.
Fabricated and misattributed citations surface before a grade is awarded.
What took hours per thesis takes minutes.
Every verdict backed by a concrete source passage — robust on appeal.
The same bar across every supervisor.
Doesn't exist. Acurio finds no match.
Source exists but says something else.
Source supports a weaker claim than asserted.
Source has been retracted / corrected.
Early language support has measurable effects on later educational success. Children with targeted support consistently achieve significantly higher completion rates (Weber 2021, p. 213). This association holds even after controlling for socioeconomic factors.
“…consistently achieve significantly higher completion rates.”
“In two of the four cohorts there was a positive but not consistently significant association.”
The source supports a weaker claim than the text asserts — overstatement.
Anonymized example · not a real thesis
Acurio is built as a lab instrument, not a data vacuum. Content never leaves the European legal area and never flows into model training.
Hosted in the EU, GDPR-compliant.
Uploaded theses and sources are never used to train models.
Content is bound to your institution account.
Run Acurio against a real submission batch from your department — with actual theses and sources — before you commit.
Book a pilot with a real submission batch from your department.