Restructure a long LLM response fast
Models often produce good paragraphs in a poor order — the conclusion lands first, examples come before the definition, or two related points sit far apart. This tool splits a pasted response into paragraph cards you can shuffle up and down until the flow reads well, then copies the result back out as clean markdown. No retyping, no losing wording.
How it works
When you paste text, the tool splits it on blank lines, treating each block as one paragraph and preserving any line breaks inside it. Each paragraph becomes a card with controls to move it up, move it down, or remove it. Reordering only rearranges whole blocks — the wording inside every paragraph stays byte-for-byte identical to what you pasted. When you’re happy with the order, copy the joined text back out, with a blank line inserted between paragraphs so it renders correctly as markdown.
Tips and notes
- Lead with the answer. For most LLM output, moving the conclusion or direct answer to the top improves readability immediately.
- Group related paragraphs together before fine-tuning their internal order.
- Remove filler paragraphs (“In summary”, throat-clearing intros) with the remove control rather than editing them out by hand.
- The split is on blank lines only, so a single hard wrap inside a paragraph won’t break it into two cards.