Every measurement the engine makes is free, for as many manuscripts as you like. The paid plans add Book Intelligence: editorial AI that explains the why behind the numbers, with a daily allowance and one full analysis run a day.
Free
$0forever
The complete manuscript diagnostic, in your browser, with cloud saving in your account.
Dialogue, show versus tell, line and copy editing, grammar
Every score shows what it was built from; nothing unmeasured is scored
Cloud library and version history
Optional LanguageTool grammar pass
Blurb material and Word export
Twenty genres, fiction and nonfiction
No credit card. Try the workspace first: open the demo.
Most chosenStarter
$5per month
For a writer in revision who wants the editorial reading alongside the numbers.
Everything in Free
Book Intelligence: deep critique, editing roadmap, opening analysis, chapter breakdown, comparable titles, query letter, beta-reader panel, market readiness
Ask your book, grounded in your chapters
Fix and rewrite suggestions on flagged passages
25 AI requests a day
One full AI analysis run every 24 hours
Fast models for every feature.
Premium
$15per month
For a writer on deadline: more room, and the deepest reading on the strongest model.
Everything in Starter
Deep critique, opening analysis and chapter breakdown read by Claude
75 AI requests a day
One full AI analysis run every 24 hours
Cancel any time from your profile.
HOW THE ALLOWANCE WORKS
Advice is advisory. The numbers stay yours.
One analysis run a day
The whole-manuscript reading (critique, roadmap, opening, breakdown, comparables, query letter, readers, market) runs together, once every 24 hours. Results stay open until the next run.
Requests for the rest
Ask your book, fixes and rewrites are counted per request against the daily allowance, 25 on Starter and 75 on Premium.
Scores never move
AI reading is marked as interpretation. It can explain a score or disagree with it; it cannot change one.
What leaves your browser
The diagnostic runs locally. AI features send the relevant passages to the provider when you choose to run them, and only then.
QUESTIONS
Before you decide.
No, and we won’t pretend it does. AuthorScrolls tells you where your manuscript is built unevenly, where the prose is getting in the way, and what a reader of your genre comes for. A line editor and a proofreader are still part of the path to publication. What it saves you is paying someone to find the problems you could have caught yourself.
The diagnostic engine runs in your browser. Your manuscripts are saved to your account, encrypted at rest. AI features send the relevant passages to the model provider only when you run them; on Starter that is OpenAI’s fast models, and on Premium the deepest features are read by Claude.
Requests go through our server with your account, never with a key in the browser. We send only the passages a feature needs, and we do not share your work with anyone else. The providers’ own data terms apply to what is sent; see our privacy policy for the current list.
A full analysis reads across your whole manuscript in several requests. One run a day keeps the allowance honest and, in our experience, matches how revision works: read the findings, revise, come back tomorrow.
Any time, from your profile. You drop to Free at the end of the billing period and keep your manuscripts, your versions and every measurement.
Word (.docx), PDF and plain text, up to 10 MB each.
Start with the numbers. Add the reading when you need it.