Prompts for headlines and hooks

What it is good at

• Generates genuine variety by requesting distinct headline formulas rather than ten similar options.
• Covers the four formats that matter most: ads, blog titles, subject lines, video hooks.
• Includes a self-ranking prompt so you're not choosing blind between similar-sounding lines.
• Fast enough to run fresh rounds when the first batch feels flat.

Where it falls short

• It doesn't know your actual open or click-through rates, so ranking is based on general principles, not your data.
• Number and stat-based hooks need a real number from you, it will invent one if you don't supply it.
• Curiosity-gap formulas can tip into clickbait; always check the headline delivers what it promises.

How to use it for creative work

1. 01

   Pick the format before writing the prompt

   Ad headline, blog title, email subject and video hook all have different length and tone rules, always specify which you need.

2. 02

   Ask for named formulas, not just 'catchy'

   Request specific approaches, curiosity gap, number-led, contrarian, benefit-led, so the ten options are actually different from each other.

3. 03

   Give it the real number or stat

   Supply any statistic, price or timeframe yourself; never let the tool invent a number for a headline.

4. 04

   Run a self-ranking pass

   Once you have ten options, ask it to rank them by likely click-through and explain the reasoning, this surfaces the strongest two or three fast.

5. 05

   Check the promise matches the content

   Read the winning headline against the actual piece and cut anything it promises that the content doesn't deliver.

Ready-to-use prompts

Write 10 headlines for [content/product/ad], two each using these formulas: curiosity gap, number-led, benefit-led, contrarian/against-the-grain, and direct/plain. Label which formula each one uses. Topic: [topic]. Audience: [audience].

Write 8 blog post titles for an article about [topic], aimed at [audience]. Mix how-to, listicle and question formats. Under 65 characters each so they don't truncate in search results.

Write 6 email subject lines (under 45 characters) for a campaign about [topic/offer], each paired with a preview text line (under 60 characters) that adds new information rather than repeating the subject.

Write 5 opening lines for a video about [topic], each designed to be said out loud in the first 3 seconds to stop someone scrolling. No 'in this video' framing, start mid-thought or mid-claim.

Write 6 ad headlines (under 30 characters) for [product/offer] targeting [audience]. Half should lead with a benefit, half with a specific number or timeframe: [supply the real number/timeframe].

Write 5 headline options that challenge a common belief about [topic]. State the common belief first, then the headline that pushes against it. Keep it defensible, not clickbait, the content must actually support the claim.

Rewrite this headline to have more tension or specificity without becoming clickbait: [paste headline]. Give 3 versions, each a different length.

Rank these 10 headlines by likely click-through for [audience/platform]: [list]. For the top 3, explain in one sentence each why they'd outperform the rest.

Write 5 headlines built around this real statistic: [stat/source]. Each should state the number early and make its relevance to [audience] obvious within the headline itself, not just implied.

Give me two genuinely different headline directions for [content], not just reworded versions of each other, so I can A/B test which resonates: one emotional, one logical/benefit-led.

Questions people ask

How many headline options should I generate before choosing?

Ten across distinct formulas is usually enough, beyond that, options start repeating the same underlying idea in different words.

Will these headlines be clickbait?

Only if you let the curiosity-gap formula overpromise, always check the final headline against the actual content before publishing.

Can it tell me which headline will perform best?

It can rank by general principles (specificity, tension, clarity) but has no access to your actual audience data, so treat rankings as a starting hypothesis to A/B test.