Frequently Asked Questions

Grammar tools work at the sentence level: comma placement, word choice, passive voice. Manusights works at the manuscript level: scope fit, methodology gaps, claims your data doesn't support, citations that don't say what you say they say, figures that need redesign. We don't fix your sentences. We tell you what reviewers will reject. Run a grammar tool after Manusights, not instead of it.

General-purpose models will tell you your paper is great. They were trained to be agreeable. Manusights runs the diagnostic rubric our 35+ reviewer network co-developed from real peer reviews at top-tier journals: the same failure patterns real editors flag in triage. Across hundreds of structured calls per review, not a single prompt. Researchers tell us repeatedly that ChatGPT confirmation bias is the thing we don't do.

Free Readiness Scan: 1-2 minutes. Full Review: about 30 minutes from upload to .docx report. Most papers deliver faster than that; we set the expectation at 30 minutes because under the hood we run 2+ hours of parallel compute and hundreds of frontier LLM calls per review. Expert Review (real human reviewer, premium tier): 3-7 days depending on manuscript length and field.

Pre-submission feedback is standard practice in academia. Every researcher circulates drafts to colleagues, presents at lab meetings, and seeks input before submission. Manusights gives you that same kind of structured editorial scrutiny, trained on the real peer-review work of 35+ top-tier-journal reviewers, before you submit. Your ideas and your work remain entirely your own.

Colleagues are valuable but have limitations: they know your work too well, may not have time for detailed feedback, and may not have experience with your target journal. Manusights is trained on the real peer-review work of 35+ active top-tier-journal reviewers, so it applies their diagnostic patterns to your manuscript with fresh eyes and structured methodology. It will catch things your lab missed because the rubric is built around the failure modes journal reviewers cite most often.

Two tiers. The Full Review is $39 (your local-currency equivalent at checkout) and delivers in about 30 minutes. The Expert Review (a real human reviewer in your field) is $1,000 to $1,800 for single manuscripts and takes 3-7 days. Most authors start with the free preview and choose from there.

For the Expert Review tier: most manuscripts are reviewed by one expert reviewer, though you can request up to three for more thorough feedback. The Full Review is automated end-to-end and runs the same diagnostic criteria the reviewer network co-developed.

The free preview is honest about this: it flags early-stage drafts and tells you what needs to be sharper before a full review is worth it, rather than forcing one. For the Expert Review tier (premium human reviewers), we assess readiness in an initial conversation and will tell you directly if it needs more development before formal review. We'd rather you succeed than waste your investment.

Manusights applies the same caliber of critique you'd receive from journal peer reviewers, but before submission, when you can still address weaknesses. Its engine is trained on the real peer-review work of 35+ active top-tier-journal reviewers, so it surfaces the blind spots, methodology gaps, and narrative issues that become rejection reasons. (The Expert Review tier puts a real human reviewer in your field on the manuscript directly.)

Yes. If your manuscript goes through journal review and requires revisions, we offer follow-up support to help you respond to reviewer comments. Many authors find this even more valuable than the initial review. We help you understand exactly what the reviewers are asking for.

Yes. The Full Review runs through Anthropic's Zero Data Retention API tier: your manuscript is processed in memory and never stored or used to train any model. For the Expert Review tier, all reviewers operate under formal NDAs and only your assigned reviewer sees your work.

The Full Review accepts all manuscripts through the standard preview flow. The Expert Review tier is scoped: we accept based on field, scientific readiness, and reviewer availability. If an Expert Review isn't a fit, we'll say so upfront.

For the Expert Review tier (premium human reviewers): yes, when you submit you can specify the expertise areas most important for your manuscript, and we'll do our best to match you with a reviewer whose background aligns. The Full Review (AI) automatically applies the field-appropriate diagnostic criteria the reviewer network co-developed.

The review is meant to strengthen your work, not dictate changes, and you retain full control over your manuscript. For the Full Review, reply to your delivery email if you want help interpreting a specific point. For the Expert Review tier, we can arrange clarification with your reviewer or discuss alternative perspectives.

Yes, we work with researchers globally. Our reviewers are experienced with manuscripts from non-native English speakers and provide feedback on both scientific content and presentation clarity. We focus on helping you meet the conventions these journals expect: framing, structure, and how evidence is presented.

Yes. Every Full Review runs vision-based parsing on every figure and every image-embedded equation in your manuscript. If your argument depends on Figure 3, we check Figure 3, including equations rendered as images, which most language-only tools skip entirely.

We verify every citation against live scholarly databases (CrossRef, PubMed, OpenAlex, Semantic Scholar, bioRxiv, medRxiv) covering more than 500 million works. We flag fabricated DOIs, mismatched titles, and citations that don't support the claim they're attached to. Bibliography formatting is something other tools handle better; we don't, and we don't try to.

We score your manuscript against the desk-reject patterns of your target journal: scope fit, structural fit, statistical reporting, and the specific failure modes editors at that journal cite most often. The score is journal-specific, not a generic rating of whether the paper is good.

We don't rewrite your sentences for you. We don't handle bibliography formatting. We don't check grammar at the comma level. If you need in-document language polishing, a writing assistant is the right tool. Run it after Manusights, not instead of. We're built for the substantive review that happens before submission, not the copyediting that happens after.