BYTETOOLS

Dialogue Tag Analyzer

Analyse how your dialogue is attributed — said and asked versus exotic tags, adverb-propped tags, action beats and long runs with no attribution at all.

Paste prose containing dialogue to see how your speech is attributed. Straight and curly quotation marks are both recognised.

Attribution is detected by looking at the words immediately before and after each quotation, so quoted titles, scare quotes and nested quotations can be miscounted. There is no correct ratio of “said” to everything else — most editors simply prefer “said” to disappear and the dialogue itself to carry the tone. Your text never leaves the browser.

What is the Dialogue Tag Analyzer?

Dialogue Tag Analyzer finds every quoted passage in your prose, with straight or curly quotation marks, and works out how each one is attributed.

  • Recognises straight and curly quotation marks, double or single
  • Splits attribution into plain tags, exotic tags, action-as-tag, action beats and none
  • Frequency table of every speech tag you actually used, colour-coded by type
  • Adverb-modified tags flagged with the adverb named
  • Unattributed runs reported with the quotation numbers involved
  • Copyable plain-text report of the whole analysis

How to use the Dialogue Tag Analyzer

  1. 1

    Paste prose containing dialogue into the text box.

  2. 2

    Choose double or single quotation marks to match your house style.

  3. 3

    Set how many consecutive unattributed lines should be flagged as a run.

  4. 4

    Read the attribution mix bars and the tag frequency table.

  5. 5

    Work through the flagged tags and unattributed runs.

About the Dialogue Tag Analyzer

Dialogue Tag Analyzer finds every quoted passage in your prose, with straight or curly quotation marks, and works out how each one is attributed. It separates plain tags (said, asked) from exotic ones (muttered, exclaimed), flags action verbs used as speech tags — laughed, smiled, nodded — and spots tags propped up with an -ly adverb.

It also counts action beats used in place of a tag and finds runs of consecutive quotations with no attribution at all, which is where readers start losing track of who is speaking. You get an attribution mix, a frequency table of every tag you used, and the flagged lines in context.

Attribution is detected from the words immediately around each quotation, so quoted titles and nested quotations can be miscounted. Everything runs in your browser and nothing is uploaded.

Frequently asked questions

What percentage of dialogue tags should be “said”?

There is no correct figure. The common editorial advice is that “said” is invisible to readers and lets the dialogue carry the tone, so most published fiction leans on it heavily. What matters more is whether the exotic tags you do use are earning their place.

Why are laughed, smiled and nodded flagged?

Because they are not ways of producing speech — you cannot smile a sentence. Editors usually ask for them to be split into a separate action beat: “I know.” She smiled. rather than “I know,” she smiled.

What is an action beat?

A small piece of action next to a line of dialogue that identifies the speaker without a tag: “I know.” She put the cup down. Beats attribute and characterise at the same time, which is why heavy dialogue scenes often use them instead of tags.

How many unattributed lines in a row are too many?

Two speakers can trade several lines cleanly before readers lose the thread; three or four is where most editors start asking for an anchor. The threshold is editable so you can match your own tolerance.

Are adverbs in dialogue tags always wrong?

No, but they are often a sign that the dialogue itself is not doing the work — “I hate you,” she said angrily. The tool names the adverb so you can decide whether it adds anything the line does not already carry.

Related tools