Manusights is a hosted AI service that reviews academic manuscripts before submission. The service is invoked at a single point in the manuscript lifecycle: after the draft is internally cleaned but before it is submitted to a journal or posted publicly. The system performs three primary analyses: Citation verification against live scholarly databases (CrossRef, PubMed, OpenAlex, Semantic Scholar, bioRxiv, medRxiv) covering more than 500 million works. Confidence labels are returned for existence, retraction status, DOI accuracy, and claim alignment. Vision-based parsing of figures, micrographs, gels, plots, and image-embedded equations. Common reviewer-objection patterns (resolution, axis labels, scale bars, claim-figure alignment) are flagged before submission. Journal-specific desk-reject scoring against a target venue's editorial triage patterns. Scoring is calibrated to the journal because what survives triage at PLOS ONE differs materially from Cell, NEJM, or Nature. Live service: https://manusights.comCitation guidance: https://manusights.com/citeData handling and security: https://manusights.com/security This deposit contains citation metadata, a public methodology summary, and the deposit license; it does not contain the Manusights service source code, which is proprietary. The contents of this deposit are released under the Creative Commons Attribution 4.0 International License (CC-BY-4.0) so that researchers may freely quote and cite this material. Cite the concept DOI (10.5281/zenodo.20072010) when referring to Manusights in general; cite the version DOI when reproducibility matters.

If you used Manusights to review your manuscript before submission, please cite this software using the metadata below. Cite the version DOI for reproducibility; link the concept DOI when referring to Manusights in general.