ManuscriptMeter

Methodology 2026.08.1

What the report checks—and where it stops.

ManuscriptMeter applies deterministic document rules and public reference metadata. It does not judge the science or rewrite the manuscript.

Document checks

We extract ordered text from DOCX files and selectable text from PDFs. Rules look for heading-level gaps, numbering discontinuities, figure and table citation order, undefined abbreviation candidates, unmatched in-text citations and references, and declaration labels.

Reference checks

When a DOI is present, ManuscriptMeter queries the Crossref REST API for resolution, title and author metadata, and available update relationships. Retraction data is sourced through Crossref’s integration of the Retraction Watch database.

Coverage is not universal. A missing update record is not proof that an article has never been corrected, withdrawn, or retracted. If Crossref is unavailable, the report is marked Partial and local document checks remain visible.

Finding levels

LevelMeaning
FixA deterministic inconsistency or unresolved reference that should be addressed.
ReviewA likely issue requiring author or journal-specific judgment.
PassThe named check found no issue; it is not a guarantee of compliance.

Explicit limits

ManuscriptMeter does not perform OCR, plagiarism detection, AI-writing detection, statistical review, scientific-validity review, authorship verification, generative rewriting, peer review, or journal acceptance prediction. Journal instructions and human review remain authoritative.

Rules version 2026.08.1 · reviewed August 16, 2026