Sentiment Analysis Tool
Score the positive or negative tone of any text using a word-valence lexicon.
What is the Sentiment Analysis Tool?
This tool scores whether a piece of text reads as positive, negative or neutral using a word-valence lexicon — a list of thousands of common words, each pre-scored from -5 (very negative) to +5 (very positive) by human raters. This is the same lexicon-based approach used by the widely-cited AFINN word list and the popular sentiment npm library, not a neural network or large language model.
How the score is calculated
The tool splits your text into words, looks each one up in the lexicon, and sums the scores of every matching word for the overall Raw Score. It also accounts for simple negation — "not good" flips the score of "good" instead of counting it as positive. The Score / Word (comparative score) divides the raw score by the total word count, which makes it possible to fairly compare sentiment across texts of very different lengths.
Why this approach, and its honest limits
A lexicon lookup runs instantly with no model to download, and its word-by-word breakdown (shown in the table below) is fully transparent — you can see exactly which words drove the score. The tradeoff is that it reads words individually rather than deeply understanding context, so it can be fooled by sarcasm, idioms, or sentence structures a full language model would catch (e.g. "this is not bad at all" may not score as cleanly positive as a human reader would judge it). Treat the result as a useful, explainable signal for things like scanning customer reviews or social comments in bulk, not a certified emotional analysis.
How to use it
- Paste or type any text.
- The sentiment score and per-word breakdown update automatically.
Everything runs locally using a bundled word list — nothing you type is ever uploaded anywhere.
Want the "why" behind this tool? Read What a Lexicon-Based Sentiment Score Actually Measures.