MetaScope. ScopeTools

Methodological audit Auditors Premium

What the methodological reviewer checks first — before they do

The CRD is cited, but the databases changed, the search date moved and the PRISMA numbers don't add up. That is a resubmission, not a typo.

You give

The manuscript (.docx, .md or .txt); ideally the PROSPERO record export (.pdf, .txt or .docx) or at least the CRD; optionally the file with the search strings.

You get

A report in .docx and .md: verdict on the first line, the extracted facts, the PROSPERO comparison, the findings table by severity (blocking, major, minor) with line and excerpt, and the reviewer statement.

How it works

# Methodological audit
$ scopetools run audit-method
Model-assisted, with deterministic checks.
Deterministic rules extract the facts from the manuscript and check the PRISMA flow arithmetic line by line. With the PROSPERO export, it compares CRD, databases, dates, study types and main outcomes field by field. Only then does a zero-context reviewer, who receives the facts and findings only (never the manuscript), write the statement.

Deterministic rules extract the facts from the manuscript and check the PRISMA arithmetic line by line; with the PROSPERO export, they compare CRD, databases, dates, study types and main outcomes field by field. Only then a zero-context reviewer, who sees the facts and the findings and never the manuscript, writes the statement.

What it costs

20 credits per manuscript (US$1.00), for subscribers. Running it again after fixing costs another 20.

Subscribers only — charged in credits per run, like every other tool.

Charged only when it delivers — a failed run costs nothing.

Credits are shared across every MetaScope tool — buy once, spend anywhere.

Limits

Reads numbers in digits; "twelve studies" in words is not read. It checks what rules can check; the science of the design belongs to the manuscript audit.

Frequently asked questions

Do I need the PROSPERO export?
Ideally yes (.pdf, .txt or .docx). With only the CRD, the record comparison is skipped and the PRISMA and reporting checks still run.
What counts as NOT COMPLIANT?
Any blocking or major finding: a PRISMA flow that doesn't add up, a database or design that differs from the record, an undeclared protocol deviation.
Does the model read my manuscript?
No. The reviewer receives only the extracted facts and the findings; the text stays in your browser.

Methods hold? Now check that the numbers reproduce with the Statistical audit.

Gabi runs it with youUpload the file, get the deliverable back.Charged only when it delivers — a failed run costs nothing.
Run this tool