Keyword Extractor
Paste any text and pull out its most important keywords and phrases. Uses term-frequency scoring with stopwords removed, and boosts multi-word phrases that look like real concepts. Handy for content tagging, SEO metadata, and pre-filtering text before computing embeddings.
15
Keywords (click a chip to copy it)
Comma-separated list
How scoring works: the text is tokenized, common English stopwords are dropped, and each remaining word is scored by how often it appears. Two- and three-word phrases whose words are all content words get a boost (2.2x and 3x) because phrases like "vector database" carry more meaning than either word alone; phrases must appear at least twice to qualify. This is a fast statistical approximation, not semantic understanding: it works best on English prose of a paragraph or more, and rare-but-important terms can rank lower than frequent generic ones.