Built for integrity decisions that stand up to scrutiny
iOriginally is a plagiarism and AI-writing detection platform for institutions worldwide. We build it around one conviction: every number an integrity office acts on should be traceable to evidence a committee — and a student — can examine.
Built for the moment a decision is questioned
An integrity finding doesn't end at a percentage. It lands in front of a student, a committee, sometimes an appeal panel — and at that moment, "the software said so" is not an answer. We built iOriginally so the answer is the evidence itself.
- Evidence on the original layout. Findings are annotated on the document exactly as it was submitted — the report a committee reads matches the work a student wrote.
- Filters that recompute in the open. Exclude quotes, the bibliography or a source and the score updates transparently. Nothing is baked in behind the scenes.
- Certificates anyone can verify. QR-secured integrity certificates resolve to public verification pages — the claim and the proof travel together.
One team, one system — engines included
We didn't license a black box and wrap a website around it. Similarity detection, AI-writing analysis and reporting are engineered together, in-house — so what the engine finds is exactly what the report shows.
Similarity engine
Web and open scholarly sources, your institution's own repository, student-to-student collusion matching, paraphrase-aware semantic matching — and anti-masking that flags homoglyph and invisible-character tricks instead of being fooled by them.
AI-writing analysis
Sentence- and segment-level probabilities, always shown separately from similarity so the two are never conflated — with a policy note printed on every AI screen and PDF reminding readers it is an aid for human judgment, not proof of misconduct.
Reports & certificates
Original View reports annotated on the exact document layout, in-browser and PDF in parity, plus QR-secured certificates with public verification pages — generated by the same system that found the matches, so nothing is lost in translation.
Every number is computed from evidence
The fastest way to lose an institution's trust is a figure that can't be explained. So we hold every number — in reports, and on this website — to the same standard: computed from evidence, or not shown at all.
No number without an engine run
Every similarity percentage and AI signal on screen comes from a real analysis of a real document. When no engine has run, the platform says so plainly — it never displays an invented figure to fill a gap.
Scores that recompute in the open
Exclude quotes, the bibliography, small matches or an entire source, and the score updates in front of you. A reader can see how the number was made — not just the result.
We publish where competitors lead
Our comparison table keeps the rows we lose. Turnitin's proprietary publisher archive is larger than our web and open-scholarly coverage — and our compare page says exactly that.
Institutions own their data
Integrity checking touches the most sensitive artifact a university handles: student work. Our answer is structural, not just contractual — run iOriginally where your policies live.
Self-host on your infrastructure
Run the entire platform — engines included — on your own servers, so student documents never leave your network. The license price is the same as managed cloud.
Never used to train models
Student work is never used to train models — ours or anyone else's. Your repository exists to serve your own matching, nothing more.
Retention on your terms
Retention windows, export and deletion sit under institution control, backed by a full audit log. GDPR-ready by design, wherever your campus is.
What we believe
Six commitments that shape every release — and every conversation with an institution.
Judgment over verdicts
Similarity and AI signals are aids for human judgment, never proof of misconduct. That framing is printed into the product — on every AI screen and PDF — not just into the marketing.
Evidence you can trace
Every number must be explainable down to the highlighted span that produced it. If a figure can't be traced back to evidence on the page, it doesn't ship.
Honesty over polish
We would rather show a losing comparison row than an impressive claim we can't back. Failed checks are never billed, and evaluation certificates are visibly watermarked.
Students are users too
A student portal with live pipeline status and receipts, because fairness includes the person being checked — not only the office doing the checking.
Own the whole problem
Engines, reports, certificates and workflow are one system, built by one team that is accountable end to end. No black boxes, no blame passed between vendors.
Institutions in control
Policy bands, roles, retention and even the deployment model are configured per institution. Your integrity policy shapes the platform — not the other way round.
Hold us to our own standard
Book a demo, or start a free 14-day evaluation and run a real batch through the pipeline — every number you see will be traceable to evidence.