The Complete AI Prompts Pack
1,080 mega-prompts for work that doesn’t sound generated. Every one tells you what it returns — and where it breaks.
$29
✓ Instant access✓ Updates land automatically✓ Works with any assistant
1,080 mega-prompts for work that doesn’t sound generated. Every one tells you what it returns — and where it breaks.
prompts
libraries
carry a stated limit
extra tools needed
Anyone can write you a prompt. Almost nobody tells you the conditions under which it stops working — so you find out on your own time.
The same prompt gives a great answer on one draft and mush on the next, and nothing explains why.
A big library is only as good as the way you search it. Without a stated use case per prompt you scroll instead of finding.
A prompt sold without its failure conditions puts the whole cost of finding them on you, one wasted run at a time.
Without an order of operations you run them in the wrong sequence and undo your own edits.
Swapping vocabulary doesn’t help. The tell is structural, and most prompts never touch structure.
If the output needs more fixing than starting from scratch, the prompt cost you time.
The last two are the product. Where a prompt has no real failure mode, it says so rather than inventing one.
[brackets] to swap01 The Rhythm Breaker "Rewrite this so no two consecutive sentences have a similar length. Include at least one sentence under five words and one over thirty. Do not change any facts. Do not smooth it out afterwards." RETURNS Prose with visible variation in sentence length. The single highest-impact change to text that reads as generated. BREAKS WHEN Under ~80 words there isn't room to vary, and it pads to comply. Use on paragraphs, not captions.
All 1,080 sit in the twelve libraries. The Prompt Builder, which writes new ones in the same format, comes with them.
Content, email, positioning, launches, measurement.
Hooks, carousels, captions, the comment-to-DM funnel.
The tells are structural, not lexical.
One-person systems, funnels, audience, sustainability.
Strategy, validation, decisions, operations.
Debugging, code review, automation.
Product copy fails by describing the object instead of the decision.
Most productivity prompts generate more tasks.
Nobody buys from a message that reads like it went to two hundred people.
Search rewards specificity too.
Learning plans, source checking, teaching.
Budgeting, pricing and margins, decisions.
Twelve libraries named after what you’re trying to do.
Ten seconds that saves you the wasted run.
Paste in your draft, your topic, your samples.
Free tier of any assistant. Nothing else to buy.
This is a tool, not a machine. It still needs you to read the output.
One payment. No subscription.
Secure checkout — your payment info is always protected.
All of them. Nothing here depends on a particular model or a paid tier — they’re instructions in plain English, not model-specific syntax.
Usually, yes — that is exactly how most large packs are built, and it is a fair thing to suspect. The check here is the “breaks when” line: it has to name a real condition under which that specific prompt stops working, and a prompt that is another prompt with a noun swapped has nothing to put there. All 1,080 carry one, every name is distinct, and every prompt states its own return. You can check it, too — it is one database, so sort it and count: 1,081 rows, being the 1,080 plus the Prompt Builder. Judge it on that rather than the number.
The condition that makes the prompt fail. For example: the rhythm prompt pads instead of varying when the text is under about 80 words. Knowing that in advance is the difference between a prompt you use twice and one you abandon.
A Notion workspace, linked immediately after purchase through Gumroad. Duplicate it into your own workspace to add notes — but bookmark the original, because that is where updates land. A duplicate is a snapshot and will not update.
Yes. No restriction on commercial use.
No. One payment, and updates are included at no extra cost.
You can keep collecting prompt packs you never open, or take 1,080 that tell you in advance which ones are going to work.
P.S. — The “breaks when” lines are the part nobody else publishes, because they are the part that cannot be mass-generated. Naming a real limit means knowing one. That’s the whole product.
Enter your email to get them!
No spam. Unsubscribe whenever.
Enter your email to get them!
No spam. Unsubscribe whenever.