Your manuscript, verified. Every finding anchored to your own words.
We audit novels for continuity errors, semantic redundancy, and AI-tells. Every finding quotes your manuscript and cites chapter and line. Continuity and redundancy findings are mechanically verified against your text, and a human editor reviews every report before it ships. Built for working novelists. Reports in 72 hours.
The problem
Everyone will help you write a novel. Almost no one will check one.
The tools market has spent its money on generation: drafting apps, co-writers, story bibles that feed prose machines. Verification got nothing. So the questions that actually decide whether a book holds up (does the timeline work, does chapter 22 contradict chapter 4, am I saying the same thing three ways) still get answered by you, at midnight, with a spreadsheet.
AI manuscript reports exist, and the cheap ones have a documented problem: they invent findings. A leading tool has been caught attributing actions to the wrong character. A report you have to fact-check yourself is worse than no report.
The reliable alternative is a human developmental editor: $1,200–$4,500 per manuscript at 2026 EFA and Reedsy rates, with a turnaround measured in weeks. Excellent work, priced for once per book rather than for every revision pass.
How it works
Order, analyze, verify. Seventy-two hours.
-
Order and send your manuscript
Pick a report, fill in the short intake — genre, POV, register, comp titles. Those answers calibrate the analysis to your book, not a generic one. Then email us your manuscript as docx or markdown.
-
Analysis, mechanical verification, human QC
We build a structured model of your manuscript and run the audit against it. On continuity and redundancy audits, every quoted excerpt is then mechanically verified against your text; quotes that don't check out never reach the main report. Then a human editor reviews the findings before anything ships.
-
Your evidence-anchored report
Within 72 hours you get findings ranked by severity, each quoting your manuscript with its chapter and line. Need it sooner? 48-hour rush for +50%.
The guarantees
- Every finding cites chapter and line. You can check any claim against your own manuscript in thirty seconds.
- If a top-severity finding is shown to be factually wrong, the report is free. Our precision is the product; we put money on it.
- Your manuscript is never used to train AI. Not by us, not through us.
- Deleted on request. Say the word and your files are gone.
FAQ
Straight answers
Is this AI?
Yes, and we won't pretend otherwise. The analysis is AI, running through pipelines we built specifically for fiction. Then every quoted excerpt is mechanically verified against your manuscript, and a human editor reviews the findings before delivery. What we guarantee is the evidence, not vibes: each finding cites chapter and line, so you never have to take our word for it.
What genres do you take?
At launch: science fiction and fantasy, LitRPG and progression, romance and romantasy, thriller, mystery, and literary fiction. That's where our calibration is deepest. Horror and anything else, ask first — "we don't serve your genre yet" costs you less than a badly calibrated report would.
What file formats do you accept?
docx or markdown. If your manuscript lives in something else, email us before you order and we'll tell you honestly whether we can handle it.
How is this different from AutoCrit, Marlowe, or ProWritingAid?
Three ways. First, evidence anchoring: every finding quotes your manuscript, and on continuity and redundancy audits every quote is mechanically checked against your text — our answer to the documented problem of AI reports inventing findings. Second, a real state model: we extract your characters, places, timeline, and world rules into a structured ledger and audit the manuscript against it, instead of asking a model to hold 100,000 words in its head at once. Third, semantic redundancy detection: existing repetition checkers are lexical — same words, flagged. We detect restatement — same content, different words. No shipping tool does that.
Will an agent or publisher object to this?
Our reports are decision support for you. Nothing is submitted anywhere, nothing is attached to your query, and nobody sees a report unless you show them.
What about refunds and turnaround?
Standard turnaround is 72 hours from the moment we confirm receipt of your manuscript; rush is 48 hours for +50%. The standing guarantee: if any top-severity finding in your report is shown to be factually wrong, the report is free. Beyond that, if something's off, email us. Every report is checked by a human before it ships, and we're small enough that "talk to us" is a real policy, not a deflection.