Marlowe and Vellam start from the same honest premise: the tool should read your book, not write it. Both are analysis-first and non-generative, so neither one puts words in your manuscript. That shared ground is real, and it is worth saying up front.
The split is in shape. Marlowe is a report you run on a finished draft. Vellam is a tool you write inside, that reads and rechecks the book as it grows. The question is whether you want a one-off diagnostic or a living workspace that tracks the whole novel with you.
Short answer: Marlowe is an analytical report tool built by literary data scientists, strong on comparable-titles benchmarking and pacing. Vellam is a stateful, analysis-first writing tool: it reads chapter by chapter inside the editor, tracks every character and thread, runs granular whole-book consistency checks, detects AI-like prose, and works in four languages. Choose Marlowe for a comps-and-market report. Choose Vellam to write the book and have it read closely the whole way.
What each tool is built around
Marlowe (from Authors A.I., co-founded by Matthew Jockers) is built around expert analysis of a finished manuscript in minutes. It is analytical, not generative, and it leans on classical machine learning and NLP rather than a large language model. Its signature strengths are comparable-titles benchmarking, marketability signals and pacing visualization. You upload a draft, you get a report.
Vellam is an analysis-first writing tool. You write in its built-in editor, and the analysis layer is the centre of the product: it reads your manuscript chapter by chapter, builds a map of every character, location and thread, checks consistency across the whole book down to specific contradictions, and tells you how AI-like the prose reads. Its AI assistant brainstorms ideas with you, but it never writes the book for you.
Functional comparison
| Capability | Marlowe | Vellam |
|---|---|---|
| Keeps every word yours (non-generative analysis) | ||
| Comparable-titles and marketability benchmarking | ||
| Pacing visualization across the manuscript | ||
| Established brand and large author base | ||
| Granular cross-chapter consistency (character, location, plot contradictions) | Vellam only | |
| Auto-built Story Atlas (per-chapter cards from your text) | Vellam only | |
| Built-in AI-writing detector (per chapter and whole book) | Vellam only | |
| Stateful chapter-by-chapter analysis, not a one-off report | ||
| Character, location and plot heatmaps across chapters | ||
| Built-in editor you can write the whole novel in | ||
| AI assistant that brainstorms ideas with you | ||
| Rule-based prose analysis, no credits (cliches, passive, weak verbs, repetition) | ||
| Inline comments and shared review with reader roles | ||
| Fully localized interface and analysis (Polish, English, German, Spanish) | ||
| One-time payment, credits never expire |
A check marks the tool built for that job. Marlowe leads on comps and marketability benchmarking and on an established brand. Vellam leads on granular consistency, a living workspace, AI detection and localization. Both keep every word yours.
The things Vellam does that a one-off report does not
It checks the whole book down to the contradiction
Marlowe is candid about its limits here. On whether you have stuck to your own rules, it says that part is down to you. Vellam takes exactly that part on. It reads your manuscript chapter by chapter, the way a reader does, keeping every prior chapter in mind, and flags four specific kinds of contradiction: a character trait, a character’s state of mind, a location description, and a plot status that stops adding up. That is the granular consistency a report leaves to you.
It builds a Story Atlas from your text
After each chapter, Vellam generates per-chapter cards for every character, location and thread, pulled from what you actually wrote, with heatmaps that show where someone disappears for a stretch. It is a living story bible that updates as you write, not a static summary you read once.
It tells you how AI-like your prose reads
Vellam has a built-in AI-writing detector. Run it on a chapter or the whole book and it scores how machine-like the writing looks, with the lowest-scoring sentences highlighted. It is the same kind of scan a publisher or contest might run, so you see it first. Marlowe does not check for this.
It lives where you write, and updates as you go
Marlowe is a discrete report on a draft you bring to it. Vellam is stateful: the analysis sits in the editor, reruns chapter by chapter, and grows with the manuscript. Some writers also find a one-off report reads generic, not specific to their individual novel. Vellam’s findings point at your actual lines and your actual characters, because they are extracted from your text every time.
An idea generator, not a prose generator
Here Marlowe and Vellam agree, and it is worth being precise about Vellam’s side. Vellam’s AI is a thinking partner, not a ghostwriter. The assistant brainstorms plot, talks through a character, and points at a line that could be sharper. What it does not do is hand you finished prose to drop into your book. There is no insert button, nothing is written into your manuscript, and the whole design assumes you write every line in your own words. Like Marlowe, Vellam keeps the book yours. Unlike a report, it stays beside you while you write it.
Built with authors and publishers, and it moves fast
Vellam is built in the open with the people who use it. There is a public feedback board where authors and publishers post requests, vote, and watch them ship, and we read and act on all of it. The roadmap is shaped by working novelists and by publishing houses at the same time, which is rare in this category, and features land quickly. The goal is the tool writers actually want, improved continuously with them.
Works in your language
Vellam’s interface is fully localized in Polish, English, German and Spanish, and its analysis is built for manuscripts in those languages. Marlowe is English-only: its classical NLP is calibrated to English, so a non-English manuscript is not really its job. If you write in Polish, German or Spanish, that difference is not cosmetic.
You stay in control
Both tools refuse to write your prose, which is the part that matters most. Vellam goes one step further on data: your text is never used to train any model, AI help is opt-in and sits in a side panel, and your manuscript is never changed for you. You see the analysis, you decide what to do with it.
Choose Marlowe if
- You want a comparable-titles and marketability report on a finished draft.
- You write in English and want a quick, expert one-off diagnostic from established literary data scientists.
Choose Vellam if
- You want to write your own novel and have the whole book read and rechecked as you go.
- You want granular consistency tracking, a Story Atlas, an AI-writing detector and sentence-level analysis in one place you write in.
- You write in Polish, German, Spanish or English and want a fully localized tool.
- You prefer a one-time payment with credits that do not expire.
Can you use both?
Yes, and they complement each other well because neither rewrites your book. Run Marlowe for a comps and marketability read on a finished draft, and write inside Vellam for the chapter-by-chapter consistency, Story Atlas, AI detection and sentence-level work before the manuscript reaches a beta reader or an editor.
Frequently asked questions
Does Vellam write prose for me?
No. Like Marlowe, it never generates your prose. Its AI is an idea generator: it brainstorms and points at lines to improve, but it never hands you finished text to paste in. You write every line in your own words.
How is Vellam different from a Marlowe report?
Marlowe is a one-off report you run on a finished draft, strong on comps and marketability. Vellam is a stateful tool you write inside: it reads chapter by chapter, tracks every character and thread, and rechecks consistency as the book grows, pointing at your actual lines.
Does Vellam check consistency the way Marlowe does not?
Yes. Marlowe says whether you have stuck to your own rules is down to you. Vellam takes that on, flagging specific character, location and plot contradictions across the whole book.
Does Vellam work for novels written in Polish, German or Spanish?
Yes. The interface and the analysis are built for Polish, English, German and Spanish. Marlowe’s classical NLP is calibrated to English, so it is English-only. For more on why this matters, see beta readers and language.