Dialogue vs Narrative Analyzer

Dialogue and narration move a story at different speeds. Conversation is fast and immediate; description and reflection slow the pace and build the world. Editors talk about the balance between them, but it is hard to judge by feel over hundreds of pages. This analyzer measures it — overall, chapter by chapter — and reviews the tags that attach speech to speakers.

How dialogue is detected

Speech is found by quotation marks. Curly quotes (“…”), straight quotes (“…”), German and Polish quotes („…“), and French guillemets («…») are recognised by default. British fiction often uses single quotes, which share their character with the apostrophe, so ‘single quotes’ are an option you switch on. Many Spanish, French and Russian novels open each speech with a dash instead; switch on dash dialogue and every line beginning with — or – counts. Everything inside the marks counts as dialogue; everything else is narrative.

The overall balance

The bar and the counters show how many words are dialogue and how many are narrative, with percentages and the number of separate speeches. Genre shapes the ideal mix: thrillers, comedy and young-adult fiction often run high, literary fiction and epic fantasy lower. There is no correct number; the value lies in comparing chapters and drafts.

Chapter by chapter

When your text contains chapter headings — lines starting with Chapter, Part or Book, or Markdown headings — a table lists every chapter with its word count and dialogue share, downloadable as CSV. A chapter at 70 % dialogue next to one at 5 % is a pacing change worth checking: is the quiet chapter deliberate, or does the story stall?

Dialogue tags

The tag table counts said, asked, replied, whispered, shouted and about thirty more, and counts tags followed directly by an -ly adverb, such as “said quietly”. The note below shows how much of your tagging uses the near-invisible “said” and “asked”. Many writing guides recommend relying on them and letting the words of the dialogue carry the emotion, saving expressive tags for moments that need them.

Reading the highlights

All detected speech is highlighted in the text, which doubles as a check: unclosed quotation marks show up as a highlight that runs too long, and speech that is not highlighted reveals a quote style you need to switch on.

A worked example

The sample has two short chapters. About a fifth of chapter 1 is dialogue, while chapter 2, with a shorter stretch of description, reaches about 30 %, and the table shows the difference at a glance. The tags list shows “said”, “replied”, “asked” and “whispered”, with two of them followed by -ly adverbs (“quietly”, “anxiously”) that a revision might cut.

Private

Manuscripts stay in your browser and are never uploaded.

Using the numbers in revision

Compare chapters against the rhythm you intend rather than against an ideal percentage. A high-dialogue chapter after a long descriptive one can feel like a release of tension; two near-identical chapters in a row may feel flat. Re-run the analysis after revising to see how the pacing curve has shifted across the manuscript.

Related tools

Count words for a submission with the Word Counter, check sentence rhythm with the Sentence Length Analyzer, cut filter words with the Weak Verb Checker, or review readability with the Readability Checker.