Clean a keyword list or caption by dropping words that already appeared once. On a platform that caps your tags, every repeat is a slot you are wasting.
In prose, a repeated word is a style question. In a tag or keyword field it is a wasted slot. Instagram accepts thirty hashtags; a marketplace listing might allow a fixed number of keywords; a game profile caps the whole field at a character count. Every duplicate in those fields is a term you could have spent on something else. This tool takes the text apart on whitespace, keeps the first appearance of each word and drops the rest, so a tag string that grew organically over six edits comes back with each term stated once.
The two modes solve different problems. The default removes every later duplicate anywhere in the text, which is what a keyword list needs. Consecutive-only removes just the back-to-back stutters (the the, and and) that survive a bad edit or a speech-to-text transcript, while leaving deliberate repetition elsewhere alone. Use the second mode on prose, because the first will quietly delete the second the in a sentence and leave you with something ungrammatical. Case folding is on by default so Fire and fire are treated as one term.
In the default mode, yes. It removes every later instance of a word, including common words a sentence needs. Switch to consecutive-only when you are cleaning prose rather than a tag list.
Yes. Splitting happens on whitespace only, so "fire," and "fire" are treated as different tokens. Strip punctuation first if you need them to match.
The first. Later copies are dropped, so the word stays at the position where you originally put it.