Can a tool really detect tone?

It can measure countable properties of a text — sentiment words, intensifiers, pronouns, hedges, formality — and that is what this analyzer does. It cannot read intent, so treat the signals as measurements rather than a verdict on how the writing comes across.

How is polarity calculated?

By counting words from a finite positive list and a finite negative list and reporting the net figure. It is a lexicon count, not a trained model, so it misses irony and negation and should be read with that in mind.

Why does my positive sentence score as negative?

Usually because it contains a negative word used positively or in negation — 'no reason to worry' counts worry as a negative hit. Word lists cannot parse negation, and the page states this limitation rather than hiding it.

What do energy and directness mean?

Energy rises with exclamation marks, intensifiers and all-capital words. Directness counts second-person pronouns and questions, which are how text addresses a reader. High energy suits promotion; high directness suits instructions.

No. Every signal is a count over a published word or punctuation list, computed in the browser. There is no model, no training data and no external service, so the same input always produces the same signals.

Tone Analyzer

Your text

Paste or type text

Tone signals

Start typing above — the results update as you write.

The Tone Analyzer reports five separate measured signals — polarity, energy, directness, confidence and register — instead of collapsing the tone of writing into a single number.

What is Tone Analyzer?

Ask what tone a piece of writing has and most answers reduce to an adjective. Tone is better understood as several properties measured at once, and the Tone Analyzer reports five of them separately so that no single verdict has to stand in for the whole.

Polarity is the first signal, and it is the most concrete. The tool counts explicit sentiment words against a finite positive and negative lexicon and reports the net figure across the text. It is honest about the method: the count is not sentiment analysis in the machine-learning sense, and a word list cannot see irony, sarcasm or negation. Not a great outcome registers as a positive hit, because the word great is on the positive list.

Energy is the second signal. It rises with exclamation marks, with intensifiers such as very, extremely and really, and with emphasised all-capital words. A high energy reading usually means emphatic or promotional prose; a reading of low means the writing is measured and declarative. For a report that is exactly right, and for a landing page it is often a weakness.

Directness is the third, and it counts second-person pronouns and question marks, both of which are how writing addresses a reader. A text that says you and asks questions reads as spoken to someone; one with no second person and no questions reads as impersonal, which is the register of policy documents and academic abstracts. Neither is better; they suit different jobs.

Confidence is the fourth, and it counts hedges: perhaps, seems, might, tends to and the rest of a published list. Dense hedging makes a claim tentative and is entirely normal in research writing, where qualification reflects the strength of the evidence. Sparse hedging reads as assertive, which can be a strength in a recommendation and a liability in a proposal.

Register is the fifth, and it reports the formality band from the same composite the Formality Level Detector uses, so the two tools agree by construction. Together the five signals describe how the tone of writing is pitched rather than naming it, and the summary sentence reads them together for reference.

The honest limits are worth repeating. The sentiment lexicon is finite; irony, sarcasm and negation defeat it. Signal counts say nothing about whether a tone is appropriate, because that depends on the audience. Use the analyzer to check consistency across a document and to spot a section that drifted into a different register, not to certify that the tone is right.

A useful routine is to compare sections rather than read the whole document at once. Paste the introduction, note the five signals, then paste the conclusion and compare. A piece of writing whose tone drifts — opening formal and closing colloquial, or beginning tentative and ending in flat assertion — usually has a structural problem underneath, such as a conclusion that claims more than the body established. The signals make that drift visible in numbers. Because every signal is a count, the comparison is reproducible from one draft to the next.

How to use Tone Analyzer

Paste your text; the five signals update as you type.
Read polarity first, remembering that it counts explicit sentiment words only.
Check energy and directness to see whether the prose is emphatic, addressed or impersonal.
Read confidence: a high hedge count makes claims tentative, which is normal in research writing.
Compare the register signal with the surrounding sections to catch a drift in formality.

When to use Tone Analyzer vs related tools

Use the tone analyzer when you want several measurable properties of the prose rather than a single label. When the question is purely how formal the register is, the Formality Level Detector reports that composite with its component weights.

For academic prose specifically, the Academic Writing Tone Checker turns the same markers into a pass-or-warn checklist, and the Sentence Variety Checker covers the rhythm that a tone reading does not touch.

Privacy & Security

This tool runs entirely in your browser — no data ever leaves your device. There is no server round-trip, no upload, no logging, and no account required. Your input is processed locally using client-side JavaScript and is never stored, transmitted, or accessible to anyone else. When you close the tab, everything disappears.