Paste a review, comment, email or a list of feedback lines to see whether it reads positive, negative or neutral. Every sentence gets a score from −1 to +1, and the words behind the score are highlighted, including negations like “not good” and boosters like “very”.
Last updated:
Neutral0.00
−1 negative0+1 positive
Highlighted words
Top positive words
Top negative words
Sentiment by sentence
Sentence scores
Advertisement
Advertisement
How to use the Sentiment Analyzer
Paste your text: a review, an email, social media comments or one piece of feedback per line.
Read the overall result: the average sentence score, from −1 (very negative) to +1 (very positive).
Check the highlights: green words push the score up, red words pull it down. Hover a word to see its value and any rule applied.
Scan the sentence table and chart to find exactly where the tone changes.
Copy the report or download CSV to keep the scores or compare many reviews.
How the sentiment score is calculated
Split into sentences. Text is split at . ! ? and at line breaks, so one review per line is scored separately.
Look up each word in the built-in lexicon (−5 to +5). Simple endings such as -s, -ed, -ing and -ly are removed when needed, so “annoying” and “annoyed” both match “annoy”.
Apply the rules: a negation within three words flips the value and multiplies it by 0.74; a booster such as “very” multiplies it by about 1.3 and a softener such as “slightly” by about 0.6; words before “but” count half and words after it 1.5×; a word in CAPS inside normal text counts 1.25×; each “!” (up to four) adds 0.36 in the direction of the sentence.
Normalize the sentence total S to a compound score with S ÷ √(S² + 15), which always lies between −1 and +1.
Average the sentence scores for the overall result.
Worked example
Sentence
Words found
Total S
Compound
The food was good.
good +2
2
2 ÷ √19 = +0.46 (positive)
The food was very good!
good +2 × 1.3, plus 0.36 for “!”
2.96
+0.61 (very positive)
The food was not good.
good +2 × −0.74
−1.48
−0.36 (negative)
Slow service, but the staff were lovely.
slow −1 × 0.5, lovely +3 × 1.5
4.0
+0.72 (very positive)
Score ranges
Compound score
Label
+0.50 to +1.00
Very positive
+0.05 to +0.49
Positive
−0.04 to +0.04
Neutral
−0.49 to −0.05
Negative
−1.00 to −0.50
Very negative
Lexicon methods are fast, transparent and private, but they cannot understand sarcasm, context or domain jargon (“this phone is sick” or “unpredictable plot” versus “unpredictable brakes”). Use the scores to sort and skim feedback, and read the important items yourself.
Frequently asked questions
How does this sentiment analyzer work?
It is lexicon-based: each known word has a value from −5 to +5. The values in a sentence are adjusted by rules for negation, intensifiers, softeners, “but”, capital letters and exclamation marks, added up and normalized to −1…+1 with score ÷ √(score² + 15), the same normalization used by the VADER method.
What do the scores mean?
A sentence scoring +0.05 or higher is positive, −0.05 or lower is negative, and anything in between is neutral. Scores beyond ±0.5 are labeled very positive or very negative. The overall result is the average of all sentence scores.
How are negations like “not good” handled?
If not, never, no, without or a word ending in n't appears up to three words before a sentiment word, that word's value is flipped and reduced to 74% (so “not good” is mildly negative rather than strongly negative). “No problem” and similar phrases are treated as positive.
Can it detect sarcasm?
No. Like most lexicon-based tools it reads words literally, so “Great, it broke again” scores partly positive. Sarcasm, irony and domain-specific meanings need a human reader or a trained model.
Can I analyze many reviews at once?
Yes. Put one review per line: each line is scored separately in the sentence table, and you can download all scores as CSV for a spreadsheet.
Is my text sent to a server?
No. The lexicon and rules are part of the page, so your text is analyzed on your device and never uploaded.
Does it work in other languages?
The word list is English only. Text in other languages will mostly come out neutral, except for emoji and emoticons, which are scored in any language.
Are all the tools on FreeToolCabin tested for quality and accuracy?
Yes! Every tool on our site undergoes automated Q/A testing in real browser environments to ensure high precision, correct output, and zero errors. If you happen to spot a bug or have a suggestion, please contact us at [email protected] and our team will review it right away.