Sentiment Analyzer

📖 Tool Introduction

The Sentiment Analyzer is a Chinese text sentiment tool that runs entirely inside your browser. It does not call any remote large language model. Instead, it scores a text with a built-in sentiment lexicon. The lexicon contains a list of common positive words, a list of common negative words, and a set of negation markers such as not, no and without. While scanning, the engine performs a longest-match pass over the input, locates every emotional word one by one, and then looks up to two characters before each hit for a negation marker. When a negation is found, the polarity of that word is automatically flipped: a phrase like not good contributes as a negative hit, while a phrase like not ugly contributes as a positive hit. After scanning, the tool counts the positive and negative hits, normalizes them into a score between minus one hundred and plus one hundred, and assigns one of five labels: very positive, positive, neutral, negative and very negative. The result is shown as a large score gauge with a slider that moves left or right along a red-to-green bar. The original text is rendered below with positive words highlighted in green and negative words in red, and two side panels list every matched positive word and every matched negative word so you can double-check the reasoning. A one-click copy exports a compact text report containing the score, the label and the hit counts. This makes the tool handy for quickly gauging the mood of product reviews, user feedback or short questionnaire responses, and because it runs fully offline it also works reliably on restricted networks. The built-in sample mixes praise with a mild complaint so you can see both the green and red highlights at once, and watch how a balanced text lands near the middle of the gauge. Each matched word is click-free and simply displayed as a colored chip, grouped by polarity, so you never have to wonder which tokens pushed the score in either direction. The whole scoring step is synchronous and near-instant, which means long reviews still return a result without any waiting spinner.

✨ Key Features

• Lexicon driven: built-in positive, negative and negation lists keep every result explainable
• Negation flip: not, no and similar markers automatically reverse the adjacent word polarity
• Five labels: from very positive to very negative, on a minus-to-plus score gauge
• Sentiment highlight: positive words in green, negative words in red inside the text
• Word lists: matched positive and negative words are summarized separately for review
• Fully local: your text never leaves the browser, with copy and download support

📝 How to Use

1. Paste the review, comment or note you want to judge into the input area
2. Click the Analyze button, or press Ctrl+Enter
3. Read the sentiment gauge and the tendency label at the top
4. Inspect the highlighted text below: green is positive, red is negative
5. Click Copy or Download to save a short text report

⚠️ Notes

• Lexicon scope: the built-in list covers common colloquial words; rare rhetoric may be missed
• Negation window: negation only flips a word within two characters before it
• Not deep understanding: sarcasm and puns are outside the scope of detection
• Mixed languages: the lexicon is mainly Chinese, English support is limited
• Reference only: the score is statistical; combine it with human judgment for decisions