Chapter Length Checker: How to Spot the Chapters That Are Too Long, Too Short, or Just Feel That Way (2026)

Every novelist asks "how long should a chapter be?" sooner or later, usually at 1 a.m., looking at a chapter that has been going on for eleven pages and showing no sign of stopping.

There's a number you can look up. Then there's the question you actually have, which is whether this chapter, in this book, is the wrong length. Those need different tools. This article gives you both: a quick way to check every chapter's length against the rest of your book, and a way to find out whether a long chapter is a problem or just long.

---

The number first, since you came for it

Most commercial novels have chapters somewhere between 1,500 and 5,000 words, and a large share sit around 2,500 to 4,000. That is a habit of the market, not a rule, and genre moves it a lot.

Read that table once and then mostly ignore it. James Patterson built a career on chapters of two or three pages, and plenty of beloved fantasy chapters run past 8,000 words. A reader never measures your chapter against a genre average. They measure it against their own patience, and against the chapters on either side.

That second point is the useful one.

---

What actually makes a reader notice length

Readers rarely say "this chapter was 6,000 words". What they notice is contrast and stall.

• Contrast. If every chapter so far has been a quick 3,000 words and chapter seven suddenly runs to 9,000, the reader feels it. The rhythm they'd settled into breaks.
• Stall. A chapter can be any length if things keep happening. A 4,000-word chapter where nothing changes feels longer than an 8,000-word one where three things go wrong.

So a good chapter length check answers two separate questions:

1. Which chapters are out of proportion with the rest of my book? That's arithmetic. You can do it in ten minutes.
2. Of those, which ones actually drag? That's about tension and reader attention, and no word counter can tell you.

---

Part one: the median rule (ten minutes, any software)

Here's the arithmetic part. You need a list of word counts per chapter. In Word, put each chapter in its own heading and select its text to read the count in the status bar; in Google Docs use Tools → Word count on a selection; in Scrivener the counts are in the outliner. Paste the numbers into a spreadsheet, one row per chapter.

Then:

1. Find the median chapter length (the middle value when you sort them). Use the median rather than the average, because one monster chapter drags the average up and hides itself.
2. Divide every chapter by the median. That's its ratio.
3. Flag anything at 1.75× or above as possibly long, and anything at 0.5× or below as possibly short.
4. Add reading time. Divide words by 250 to get rough minutes for an average adult reader. Fifteen minutes is about one commute stop or one bedtime sitting; that is often where readers decide whether to start the next chapter.

The flags aren't verdicts. They're the short list of chapters worth looking at first.

---

A worked example: 24 chapters, 89,100 words

Here's an invented manuscript run through the median rule. Median chapter: 3,400 words. Average: 3,713, already pulled upward by the long ones, which is exactly why we don't use it.

Reading the table

• Chapter 7 (2.62×, 36 minutes) is the obvious one. A reader who's been getting a chapter every quarter of an hour now faces more than half an hour without a break. Often a chapter this far out is two chapters wearing one heading, with a natural cut somewhere in the middle where the scene or point of view changes.
• Chapters 13 and 14 (1.88× and 1.79×) sit back to back, in the middle of the book. Two long chapters in a row at the midpoint is the classic shape of a sagging middle. They might be fine. They're the first place to look if a beta reader says "it slowed down somewhere".
• Chapter 15 (0.32×, 4 minutes) could be a mistake or a masterstroke. A 1,100-word chapter straight after two long ones can land like a slap: a death, a reveal, a single phone call. If it's that, keep it. If it's a scene you didn't know where else to put, merge it.
• Chapters 21 to 23 shrink steadily (0.76× → 0.56×). That isn't a problem. Shortening chapters as you approach the climax is one of the oldest acceleration tricks there is, and the numbers show it working.
• Chapter 24 (1.53×) is long but unflagged, and that's typical: final chapters often hold the climax and the resolution together.

The arithmetic narrowed 24 chapters down to four questions. It can't answer any of them.

---

Part two: is the long chapter actually slow?

This is where word counts run out. To know whether chapter 7 drags, you need to know what happens inside it, scene by scene, and how a reader experiences that.

Four quick tests you can run by hand:

1. The change test. List what is different at the end of the chapter compared with the start: what someone knows, wants, has or has lost. If the list is short and the chapter is long, it drags.
2. The scene count. Count the scenes in it. One long scene can carry 8,000 words if it builds. Four scenes where the third repeats the first usually can't.
3. The cut test. Find the point where you'd cut it in two. If a strong ending already exists halfway through, you've found a chapter break you've been hiding.
4. The skim test. Hand it to a reader and ask them to mark the first paragraph they skimmed. Where they start skimming is where the chapter ended for them.

These work. They're also slow, because you have to do them for every flagged chapter, and the last one needs another person every time.

---

Why authors check chapters in StoryWith AI

We build StoryWith AI, so treat this part as a recommendation from an interested party. The median rule and the four tests above need nothing but a spreadsheet.

First, an honest note: StoryWith AI doesn't have a button labelled "chapter length checker". A word count is the easy part, and any word processor gives you one. What StoryWith AI does is answer the harder question from part two, which long chapters actually drag, across the whole book at once.

Your book is kept as chapters made of scenes. Import a .docx, .txt or .pdf manuscript, or write in the editor, and every scene carries its own word count, shown as you write. You see what a chapter is made of, not just how big it is, which is the starting point for the scene-count and cut tests.

Every scene gets a tension score from 0 to 100. The Story Map lays your chapters out as columns of scene cards, each with a tension bar, and you can drag scenes between chapters to try a new break without cutting and pasting text. A long chapter whose scenes all sit at the same low tension is the sagging-middle pattern from the table above, visible at a glance.

The arc analysis finds the flat spots for you. It reads the shape of the whole manuscript, names the overall arc, and lists flat spots (from where, to where, and why) and abrupt jumps where tension changes without preparation. If chapters 13 and 14 really are the sag, this is where you'll see it named.

Readability flags the walls of text. The Readability page marks scenes with long sentences, long paragraphs or heavy, many-syllable words, and calls out any scene that's one unbroken block. A chapter can be a reasonable length and still feel exhausting because it never gives the eye a place to rest.

Reader personas tell you where attention goes. Run a chapter past a panel of three to five AI reader personas (a thriller fan, a literary reader, a casual reader and others) and each one reports what they liked, what they struggled with, the slowest section, and whether they'd keep reading. It's the skim test without waiting for a person, and it's hedged honestly as how this type of reader may react.

Real readers show you exactly where they stopped. Feedback sessions send a private link to between 3 and 20 human readers and track how much of the text each person read, how long they spent, and the paragraphs where readers dropped off. If everyone stalls 5,000 words into chapter 7, that's your cut. StoryWith AI can then compare what the AI personas predicted with what the real readers did.

A readiness check for the book as a whole. The readiness check, which spends no AI credits, treats a draft under 20,000 words as too thin to read end to end and 50,000+ as a full draft, and flags scenes under 100 words, the stubs that usually make a chapter look shorter than it reads.

Want to try it on one chapter first?

The free Story Rater takes a chapter of up to about 5,000 words, with no account, and rates it on hook, pacing, clarity, dialogue and character, with the three fixes that would lift it most. Paste in your flagged chapter 7 and see whether pacing is where it scores lowest.

---

Fixing a chapter once you know what's wrong

That last row matters more than it looks. Plenty of manuscripts have perfectly even chapters and still drag, because every chapter is the same kind of scene at the same intensity. Even lengths can hide a flat book as well as uneven ones can expose one.

---

Questions writers ask about chapter length

How many words should chapter one be? Often a little shorter than your median, because readers (and agents) decide fast. Many openings land between 2,000 and 4,000 words. If your first chapter is your longest, check whether it's front-loading backstory. Our guide to first chapter analysis goes deeper.

Is it bad if my chapters are all different lengths? No. Variety is normal, and deliberate short chapters are a powerful tool. What's worth a look is a chapter far out of line with its neighbours for no reason you can name.

Do agents care about chapter length? They care about total word count for the genre, and about whether the opening pages pull them in. Chapter length matters to them only when it affects how the book reads.

What's a chapter length checker, exactly? At its simplest, any tool that counts words per chapter. The useful kind compares each chapter with the rest of the book and points you at the outliers; the most useful kind also tells you which of those outliers readers actually experience as slow.

Can I have a 500-word chapter? Yes, if it does one sharp thing. Very short chapters work best in contrast with longer ones around them.

How long does it take to read a chapter? Roughly words ÷ 250 minutes for an average adult reader: a 3,000-word chapter is about 12 minutes. Fast readers will be quicker; dense prose slows everyone down.

---

Your ten-minute starting point

Make the spreadsheet tonight. Find your median, divide, flag the outliers, and you'll have a short list of chapters to look at, usually three to six of them, instead of a vague feeling that "something in the middle is slow".

Then, for each flagged chapter, ask the second question: does it actually drag? Run the change test yourself, or import the manuscript into StoryWith AI and let the tension map, arc analysis and readers show you. For more on the second question, read Pacing in Writing and Story Structure Analyzer.