Peer review is broken.

Libra scores academic papers across 28 mathematical dimensions. Transparent. Reproducible. Every score public. Every decision on the record.

28
Dimensions scored
Every claim, every equation, every citation
600M+
Papers indexed
OpenAlex, Crossref, Semantic Scholar, arXiv
63K+
Retractions tracked
What peer review missed
0
Credentials checked
The math doesn't care who you are
OpenAlex Crossref Semantic Scholar arXiv PubMed Retraction Watch

28 dimensions. Every paper.

Submit a paper. Libra analyzes every equation, every citation, every statistical test, every claim. 28 dimensions of mathematical verification — from semantic consistency to proof validity to whether the conclusions actually follow from the evidence.

Results in hours. Not months.

D3
Safety
D16
Proof Engine
D22
Type Safety
D25
Privilege Esc.
D25: Claims exceed evidence

What they claim. What the data shows.

Publishers say peer review takes 6–8 weeks. The data says 3–12 months. They say their process is rigorous. 63,000+ retractions say otherwise. Libra pulls public data from Crossref, OpenAlex, and Retraction Watch and shows exactly where the system fails.

Every number backed by data. Every claim verifiable.

Publisher Claimed Review Actual Median Retractions
Elsevier 6–8 weeks 147 days 12,400+
Springer Nature 4–6 weeks 168 days 8,200+
Wiley 6–10 weeks 193 days 5,100+
Source: Crossref, Retraction Watch — public data

Public review. On the record.

Every Libra score can be challenged by anyone — with evidence. Authors respond publicly. All exchanges permanent. No anonymous gatekeeping. No private rejections. No papers vanishing because one reviewer had a bad day.

Science is a conversation. Libra makes it public.

Challenge to D3 (Safety) Open

"The p-value in Table 3 appears to be post-hoc adjusted. The original analysis plan registered on OSF specifies a different primary endpoint."

Dr. R. Chen · March 15, 2026
Author Response Resolved

"Correct — Table 3 reflects exploratory analysis. We've updated the manuscript to clearly label it as such. Primary endpoint results are in Table 2."

Authors · March 16, 2026

Who profits from free labor?

Researchers write for free. Reviewers review for free. Publishers charge universities $10K–$30K per journal per year to read the results. On a product made entirely by unpaid academics. Profit margins: 30–40%. Higher than Apple's.

$3.5B
Elsevier
$2.0B
Springer
$2.0B
Wiley
37%
Avg margin

The math doesn't care about credentials.

It cares whether the paper is true.

Score a paper View pricing