How it works
Buy a pack, download it, fill in the blanks. No apps, no subscription, no prompt engineering course.
Choose and pay
Single pack or the complete bundle. One payment through a secure checkout, no subscription, no account.
Download in seconds
Your download link arrives by email immediately. Files never expire, and updates land in the same place.
Fill in and paste
Replace the [BRACKETS] with your details, paste into your assistant, and refine from a strong first draft.
Anatomy of a prompt
Why these work when "write me an email" doesn't
A role, not a request
Naming the expertise you want changes the vocabulary, structure and standards the assistant applies before it writes a word.
Your context, clearly marked
Every variable sits in [BRACKETS], so you know exactly what to change and nothing generic slips through.
Rules with numbers in them
"Short" means nothing. "Under 120 words, no superlatives, subject line under 6 words" produces something you can actually send.
A self-critique step
Asking the assistant to name its own weakest assumption is what turns confident filler into something you can trust enough to edit.
What you actually receive
Indexed and printable, with every prompt on its own page and its variables listed at the top.
Markdown
One plain-text file: drop it into Obsidian, a snippet manager, a repo or a CLI assistant.
Notion-ready
A CSV import that builds a searchable, taggable prompt database in your workspace.
Which assistant should I use them with?
All of them, including the free versions, and they work fine on a phone. Nothing in a pack depends on a particular model, plugin or paid tier: they are plain text instructions. Longer prompts tend to do better on the larger models, and the packs note where a prompt benefits from a model that can read files or browse.