Verified against ChatGPT · 2026-07-30
Tighten writing down to a strict word limit without losing the point
Cuts a draft down to an exact word count by removing filler, hedging, and repetition first — instead of the usual AI move of chopping whole sentences and losing the actual argument.
The prompt
Ready to copy — highlighted parts are example details you can swap.
TASK: Cut the text below to 40 words or fewer, while preserving every distinct point it makes. Do not simply delete sentences from the end — actively tighten the language throughout.
TEXT TO TIGHTEN:
I think it's important to note that, at the end of the day, our Q3 results were actually pretty strong across the board. In order to achieve this, the team worked really hard and put in a lot of effort throughout the quarter, and I want to say that this hard work is what really drove the strong results we saw. Revenue grew by 18% year over year, which is a great outcome, and it's worth mentioning that this is the best quarterly growth we have had in over two years.
CUTTING PRIORITY (in this order):
1. Remove filler phrases and hedging ("it's important to note that," "in order to," "at the end of the day," "I think that")
2. Remove redundant restatement of the same point in different words
3. Combine sentences that make related points separately
4. Shorten remaining sentences by cutting unnecessary qualifiers and adjectives
5. Only as a last resort, cut a lower-priority supporting point entirely — and if you do, tell me which one and why
CONSTRAINTS:
- Keep the original tone (formal stays formal, casual stays casual)
- Keep every number, name, and specific claim from the original — don't generalize away specifics to save words
- The result must still read as complete sentences, not a bullet-point fragment version of the original
- Report the final word count at the end, and flag if you had to go even 5% over the target to keep the meaning intact rather than silently missing the target
Return the tightened version, then the word count, then any cuts made under priority 5.Customize the highlighted detailsoptional — the prompt above already works
Why this works
Most "shorten this" prompts fail the same way: the model just deletes the back half of the text or converts prose into a choppy bullet list. Giving it an explicit, ordered cutting hierarchy — filler and hedges first, redundancy second, sentence-combining third, and whole-point deletion only as a documented last resort — forces the model to tighten language throughout the piece instead of amputating from the end. Naming actual filler phrases as a mini few-shot list ("it's important to note that," "at the end of the day") is more effective than the instruction "remove filler," since those exact hedges are also the phrases models themselves default to when asked to write more content, so they need to be named to be reliably cut. The self-report constraint (report the count, flag if over target) closes the loop instead of leaving you to recount by hand — pair this with the word-counter tool to verify the model's own count, since LLM word counts are frequently off by a few words.
What you get back
Our Q3 results were strong: revenue grew 18% year over year, the best quarterly growth in over two years, driven by the team's work throughout the quarter. Word count: 27 No priority-5 cuts were needed — all points fit within the target after tightening filler and redundancy alone.
Verified against
ChatGPT GPT-5.1 · 2026-07-30
Claude Claude Sonnet 4.5 · 2026-07-30
Gemini Gemini 3 Pro · 2026-07-30
Changelog
- 2026-07-30 — Added the ordered cutting hierarchy after finding unordered instructions led models to cut whole sentences from the end rather than tighten language throughout.
- 2026-07-11 — Initial version published.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
