Unique Words

What Unique Words helps you do

List each distinct word in the order it first appears. Unlike the frequency tool, this gives you the vocabulary of a text without repeats, preserving first-seen order. It is a quick way to see which terms a document actually uses before you build a glossary, a tag list or a keyword set.

The tool runs in your browser. Your text never leaves your device.

How it works

The text is tokenized into words, punctuation is treated as a separator, and each word is lowercased to form a comparison key. The first spelling of a word is kept in the output; later occurrences are skipped. The result is a newline-separated list in first-seen order.

How to use Unique Words

1. Paste the text.
2. Generate the list of unique words.
3. Scan the list and copy the vocabulary you need.
4. Sort or format the result in the destination document.

When to use it

• Build a vocabulary list from a passage.
• Check which distinct terms a document uses.
• Prepare unique keywords before analysis.
• Find every term in a contract or policy for a review checklist.
• See whether a translation introduced new words.
• Build a tag list from a small set of articles.
• Extract the vocabulary of a short story for a lesson.

When not to use it

• Do not use it to rank importance; it does not count occurrences.
• Do not expect common words to be filtered out; stop-word removal is a separate step.
• Do not rely on it for languages without clear word boundaries.

Example

A product description repeats the words lightweight and durable. The list keeps lightweight, durable and the other distinct terms once each, in the order they first appear, so a glossary can be built from the same text. A contract with defined terms produces a short vocabulary of the parties and obligations. The output is a vocabulary, not a frequency table; sort it or filter stop words in the destination.

What you can do with Unique Words

• Extract every distinct word from pasted text.
• Keep the first-seen order rather than sorting.
• Compare words case-insensitively for deduplication.
• Ignore punctuation between words.
• Copy the whole list in one action.
• Keep the source text in the browser.

Example: build a flashcard list

Paste a short passage and generate the vocabulary. Remove words the learner already knows, keeping the list to a manageable size.

Copy the remaining terms into the flashcard tool and add definitions there. The destination can sort alphabetically even though this list follows first-seen order.

What to know before you start

• Case and punctuation handling determine whether two forms count as one word.
• It preserves first-seen order but does not sort the result.
• It does not remove stop words or rank terms by frequency.
• Hyphenated and apostrophe forms follow the tokenizer's rules, which may not match a style guide.
• Languages without spaces between words are not segmented reliably.

Common mistakes

• Expecting a frequency count; this tool lists each word once.
• Assuming the list is alphabetical when it is first-seen order.
• Forgetting that case differences are collapsed, so The and the count as one word.
• Treating hyphenated terms as two words when the tokenizer keeps them together.
• Pasting code or markup and counting identifiers as prose vocabulary.
• Treating every listed term as a definition candidate.
• Assuming the first occurrence is the preferred spelling.

Troubleshooting

If a word appears twice in the output, check for a different character such as a curly apostrophe or an accent that the tokenizer treats as part of the word.

If a hyphenated term is split, the tokenizer follows its word rules; add the term to the destination list manually if it must stay together.

If the list is very long, the source may contain boilerplate or repeated navigation text; trim it first.

If a non-Latin script looks wrong, the tokenizer may not recognize its word boundaries; use a language-aware tool.

If a term appears under two spellings, the list treats them as different words. Normalize the source or accept both forms, and remember that the tool cannot tell a product name from a common word.

If the list seems to miss a word, check whether it appeared only inside a URL, an email address or code, where the tokenizer may handle it differently.

If a term is split by a slash or plus sign, it may become two words; normalize the source if the term must stay whole.

If the list includes numbers, they are tokenized as words; filter them in the destination.

Related standards and documentation

• Unicode text segmentation (UAX #29) explains why word boundaries differ between tools.
• Punctuation is treated as a separator, not as part of the word.