Loading page
About this tool
Measure dense prose, set a compression target and bullet budget, and build the prompt that turns paragraphs into scannable points.
Prose to Bullet Prompt measures the text you paste — words, sentences, paragraphs, Flesch Reading Ease and Flesch-Kincaid grade level, all computed from their published formulas — then turns a compression percentage into a concrete bullet budget: target word count, number of bullets and words per bullet. The prompt it produces tells the model exactly how many bullets to write, in which style, and forbids dropping any fact, number or date from the source. For anyone converting a dense document into something a reader will actually scan.
Open Prose to Bullet Prompt on AltFTool — it loads instantly in your browser.
Paste or type the text you're working with.
Choose how it should be transformed or analyzed.
Copy the finished text into your document or post.
40% of 64 words is 26 words, which at 12 words a bullet is 3 bullets — no guessing.
Flesch Reading Ease and Flesch-Kincaid grade are calculated from the source text itself.
The prompt makes losing any number, name or date an explicit failure condition.
A score of 60 to 70 is considered plain English and is roughly an eighth-to-ninth grade reading level; below 30 the text is difficult and best suited to a specialist audience. The score falls as sentences get longer and words get more syllables, so shortening sentences is usually the fastest way to raise it.
Around 8 to 14 words works for most lists — long enough to carry a complete idea, short enough to scan in one fixation. If a bullet needs a semicolon or a second clause, it is really two bullets.
Keep a flat list at or under about seven items; beyond that, group them under two to four short headings. The seven-item guideline comes from Miller's 1956 paper on short-term memory span, and while the original claim is often over-applied, the practical effect on scannability is real.
It will if you compress hard — that is the point of setting an explicit target. The generated prompt requires the model to state, on a final line, anything it had to drop to hit the length, so you can decide whether to raise the compression percentage.