Softonic

Tone Analyzer

Step 1 / 3
Next

You wrote a reply to a client and you’re not sure whether it lands as polite-but-firm or accidentally snappy. This analyzer checks your text against seven small English word lists (positive, negative, formal, casual, urgent, assertive and polite, about 110 markers in total), counts the hits, and reports the sentiment, register and pressure of the message alongside punctuation and sentence-length stats. It is plain word counting, not AI, which is exactly why it is instant and predictable: a sanity check that catches the obvious tonal mismatches in two seconds, not a verdict.

How the analyzer reads your text

  1. 1

    Paste the text

    An email, a Slack message, a paragraph: anything from a sentence upward. English text only; the marker lists are English.

  2. 2

    Words are matched against the lists

    Each word is compared, as an exact whole word, with the seven lists: positive, negative, formal, casual, urgent, assertive and polite.

  3. 3

    Sentiment and register are scored

    At least two more positive hits than negative ones = Positive (and the reverse for Negative); whichever of the formal or casual lists collects more hits sets the register, with ties reading as Neutral.

  4. 4

    Read the structural signals

    You also get exclamation and question counts plus the average sentence length, which often carry tone more than single words do.

What each signal actually measures

Signal Example markers
Positive great, love, amazing, excellent, glad
Negative bad, hate, terrible, awful, sorry
Formal therefore, moreover, pursuant, hereby, shall
Casual yeah, kinda, gonna, basically, literally
Urgent urgent, asap, immediately, critical, deadline
Assertive must, need, ensure, require, will, always
Polite please, thanks, kindly, appreciate, grateful

These are excerpts from the actual lists, roughly 110 English words in total. Matching is whole-word and exact, so close variants (“thx”, “thankful”, “needed”) are not counted.

Reading the numbers

  • A single urgency word sets Urgency to High. One “asap” in a throwaway line marks the whole message as urgent, which mirrors how readers react to it.
  • More than three exclamations in a short message reads as intense, even when the words themselves are polite.
  • Questions outnumbering statements often reads as consultative or unsure, depending on the context.
  • A polite count of 0 in a request is worth a look: a simple “please” or “thanks” often softens the message a lot.
  • An average sentence above 25 words feels formal and heavy to readers, even though the register verdict itself only counts the word lists.

Tips

  • This is lexicon-based, not contextual. Sarcasm (“oh, great”) will read as positive. Re-read your own text with a human eye before sending.
  • If the register comes out as Casual and you wanted Formal, check for “just”, “basically” and “really”: they are the biggest downshifters.
  • For high-stakes messages (apologies, refunds, complaints), aim for a polite count of 2+ and an urgent count of 0 unless it is a real emergency.
  • Short messages (under 20 words) are too sparse for the register signal to mean much.

Built for English text

The marker lists are English, so the verdicts are only meaningful for English writing. Text in other languages mostly comes out Neutral because few or no markers match, and text written entirely in a non-Latin script (Arabic, Russian, Thai, Chinese, Japanese, Korean) may return no analysis at all, since the word counter only recognizes Latin letters. The punctuation stats likewise assume the Latin marks “!”, “?” and “.”.

Frequently Asked Questions

No. It counts exact whole-word matches against seven fixed English word lists and applies simple thresholds. That is what makes it instant, free and predictable; it also means it has no understanding of context.

No. The analyzer is lexicon-based and treats “oh, great” the same as “this is genuinely great”. Sarcasm detection requires context models well beyond a word list.

No. The lexicons are English, so other languages come out Neutral at best, and text in non-Latin scripts (Arabic, Russian, Thai, Chinese, Japanese, Korean) may produce no analysis at all. Use it on English text for meaningful results.

Anger often shows up in punctuation and sentence rhythm rather than in specific words. Check the exclamation and assertive counts, they pick up tone that the sentiment field misses.

As a rough filter, yes. Keep samples of your target voice and compare the signals side by side with new copy.

Yes, to us and nowhere else: the analysis runs on our server, so the text travels there with each update, is processed on the fly and is not stored. In the step-by-step view it also rides the page address between steps. Better not to paste confidential content.

Related Tools

Random Object Generator

Generate random everyday objects for drawing prompts, writing ideas, classroom games, scavenger hunts and quick creative exercises.

Headcanon Generator

Create curated headcanon prompts for fictional characters and OCs. Choose focus, tone and canon distance; generate locally in your browser tab.

Bibliography Generator

Create a reference entry for a book, journal article or web page in APA 7, MLA 9, Chicago 18 or Harvard, with italics that survive copy and paste.

Grammar Checker

Quick grammar and spelling pass for English text: fixes common typos, missing apostrophes in contractions and repeated words.

IEEE Citation Generator

Create a single IEEE-style numbered reference for a journal article, conference paper, book, or website with DOI, URL, and access-date fields.

Random Word Generator

Generate random English word lists. Choose everyday, nature, creative, or tech words, set length and count, then copy the result.

Tool available in other languages