What it is

Book Check Genie is an AI manuscript review tool for fiction authors and editors. You upload a full manuscript and it flags continuity problems, so you can fix them before readers, reviewers or clients find them.

What it checks

Six categories of continuity, in every full check:

- Character: appearance, age, relationships and facts that change without explanation

- Timeline: dates, durations and sequence of events that don't add up

- Numeric: counts, distances, quantities and other numbers that shift

- Plot state: what a character knows, has or has done, and whether the story remembers it

- Worldbuilding: rules, places and systems that contradict themselves

- POV and voice: slips in point of view or tense

It also runs polish checks: tense consistency, possible name variants, repeated phrases, and readability statistics.

How findings work

Where it can, each finding comes with a supporting quote from your manuscript and the chapter it appears in, so you can check it yourself. Quotes are verified against your text, and ones that can't be verified are dropped. Findings are queries to look at, not corrections. The AI can miss things and can be wrong, so it supports a careful read and doesn't replace one.

Useful for series and revisions

- Check a book against approved series canon, so facts stay consistent across books

- Mark each finding as resolved, intended or dismissed

- Re-check a revised draft to see which issues are fixed

For editors

Export the report as a numbered query list in Word (.docx), ready to send to an author. PDF and CSV exports are included too.

What you can upload

.docx, .pdf, .epub, .rtf, .txt and .md files.

Try it free

There's a free sample check on the first part of your manuscript, no card needed. Full-length checks and re-checks are on paid plans.

Your manuscript

Manuscript text is sent to Anthropic's Claude API to run the check. Handling, retention and deletion are explained at bookcheckgenie.com/confidentiality. You can delete a book or your whole account yourself.

Feedback wanted

I built this and I'm looking for honest feedback. Where do the findings help, where do they miss something, and where are they wrong? Tell me what you'd want it to check that it doesn't.