PeerTrust
Research Integrity Forensics
PeerTrust surfaces citation-anomaly indicators for human-led research-integrity review, on a fully reproducible evidence chain.
peertrust.report- Sector
- Research Integrity · Scholarly Publishing
- Type
- Web platform · free public beta
- Inputs
- Researcher ORCID · Venue ISSN
- Source data
- OpenAlex scholarly record
Trust in a research record is easy to assert and hard to evidence.
Integrity teams are asked to judge a researcher's standing from numbers that were never built to answer that question — and to defend whatever they conclude.
Citation counts don't answer the question
A high citation count is treated as a proxy for standing, but it says nothing about where those citations came from or who benefited from them. On its own, the number is often limited.
The patterns that matter go unexamined
Reciprocal citing, citations concentrated among past collaborators, a sudden jump in output — the signals a reviewer would want are usually not examined, because nobody has the time to trace them by hand.
A finding has to survive an appeal
An integrity decision gets challenged. If the analysis can't be reproduced from preserved sources, and the method can't be read and checked, the finding doesn't hold.
Three patterns, measured the same way every time.
A citation happens when one paper references another. PeerTrust follows those connections across a scientist's published work and highlights unusual patterns that may deserve a closer look — three named checks, each showing its own measurement.
Citation-return pattern
Traces whether citations flow back to the authors a researcher has cited — repeated reciprocal citing, measured rather than assumed.
Past-collaborator concentration
Measures how much of a researcher's incoming citation activity clusters around people they have already published with.
Publication jump
Looks at the largest one-year increase in publication output, so unusual growth is visible against a researcher's own history.
Three states. None of them a conclusion.
Every check reports triggered, at threshold, or within threshold, with the measurement behind it and the evidence chain underneath. What it never reports is a verdict.
From an ORCID to a defensible dossier.
Six stages run in the open, with live progress as each one completes, so a reviewer can see how the result was reached rather than just what it says.
Identify researcher
Confirm the identifier against the source registry.
Collect publications
Pull the publication record from OpenAlex.
Citation-return
Check citation-return patterns across the traced works.
Collaborator share
Check how far citations concentrate around past collaborators.
Citation network
Build the network so the records can be explored.
Build the report
Assemble the reviewer-grade evidence chain.
The result is only as good as
what it can prove.
Transparent metrics, preserved evidence, and reviewer-grade dossiers that stand up to appeals. Everything behind a PeerTrust result stays inspectable — long after the live record has changed.
“Every signal can be traced back to the public scholarly records behind it.”
Evidence locket
Source records preserved with content hashes, so the material behind a result stays fixed after the live record moves on.
Chain of custody
Every action logged and traceable, from the first lookup to the exported dossier.
Reproducible replay
Any result can be re-run from the stored sources and land in the same place.
Published formulas
Every measure is documented in the open — no opaque model standing between the data and the reader.
Interactive report
The result with its measurements, explorable in the browser.
Citation network
How the records in the review connect to one another.
Signal profile
The measured patterns and publication timeline, side by side.
Institution dossier
A printable PDF built to be read by a committee.
Four choices that shaped the platform.
Indicators, not verdicts
PeerTrust never decides that someone did something wrong. It surfaces indicators, marks them triggered, at threshold, or within threshold, and hands the judgment to a human reviewer. The interface says so on every screen.
No black box
Every formula is published and every measure interpretable. A tool that makes claims about a researcher's record has to be more transparent than the thing it's examining, not less.
Built to be appealed
We designed backwards from the hardest moment — a challenged finding. Preserved snapshots, content hashes, an audit trail, and replay exist so the dossier holds up months later.
Try it before you trust it
The beta runs with no login and no account. An integrity office can check a real ORCID and read a real evidence chain before ever talking to us.
Live as a free public beta, with the method in the open.
Anyone can check a researcher against the public scholarly record without an account, read the measurement behind every indicator, and export a dossier a committee can work from. Institutions that need volume and role-based access get in touch from there.
PeerTrust sits alongside the wider ScienceWerx platform we designed and built.
Read the ScienceWerx case studyWhat changed.
Outcomes
- Anyone can check a researcher against the public scholarly record without creating an account.
- The measurement behind every indicator is published, so a reviewer can audit the reasoning instead of trusting a score.
- Every signal traces back to the public records it came from, and the dossier exports for a committee to work from.
- Findings are framed as indicators for human-led review, never as verdicts.
Ready to build what's next?
Let's turn your next business challenge into a digital experience that works.