Integrity checks your editors can run themselves.

Image manipulation, leaked review copies and undisclosed AI help cost publishers retractions and trust. We do not judge whether research is right. We give your integrity team evidence it can check without asking us.

Three jobs, three kinds of evidence

Instrument images

An image sealed at capture can be shown to be unchanged since its seal. Any later edit breaks the seal. The seal says nothing about whether the experiment behind the image was sound.

Review copies

Each reviewer receives a copy with its own invisible mark. If a figure leaks, reading the mark shows which copy it came from.

AI-assisted steps

When an AI agent screens submissions or checks references, each step goes into a signed receipt, so you can later show what the agent did and when.

Legal and policy status (26 October 2026)

Disclosure rules for AI use come mainly from publisher policies and COPE guidance, not from one law. For AI-generated images, Art. 50 of the EU AI Act (Regulation (EU) 2024/1689) requires machine-readable marking by providers and disclosure of deepfakes by deployers. Check the current text and your own policy.

Not legal advice. Evidence supports an integrity decision; it does not make it.

What we prove and what we do not

Check a receipt yourself Ask for a pilot