Start here

This site is a free library of AI prompts. You find a job you need done, fill in the blanks, and copy the prompt into a chat or coding tool.

How to use this site

1. Find a prompt. Search the library or filter by category. Each card shows the model notes and the variables you will need to fill in.
2. Fill in the blanks. Anything in {{double braces}} is a slot for your own text: an error, a draft, a diff. On a prompt's page, the boxes update the text, and Copy prompt copies the filled-in version. The Copy button on a library card copies the prompt with the blanks still in it, so you replace them after pasting.
3. Paste it into an AI tool. The prompts are plain text. The model notes on each one say if it needs a coding tool, web browsing, a long context window, or a small local model.
4. Read the output. Check facts, numbers, citations and code before you rely on them.

Beginner guides

Three short guides. They cover the prompts in this library, and prompts you write yourself.

• How to use these prompts. Fill in the variables, pick the right AI tool, paste, and check the output. A two-minute walkthrough of how this library works.
• 10 habits for writing better prompts. Ten small habits that make AI output more useful, each with a free prompt from the library that shows it in action.
• Prompts for coding vs writing vs research: what changes. What context to give, what to ask for, and how to check the result when you prompt for code, for writing and for research.

10 free prompts to try first

Picked from across the free library. Each link opens the prompt page, where you can fill in the variables and copy it.

1. Spec-first app kickoff Vibe coding
   Asks for a short build spec (summary, MVP, out of scope, data model and a step plan) and does not write code until you say go.
2. One small step at a time Vibe coding
   Implements only the step you name, lists the files it will change, and tells you how to verify the result.
3. Explain this error like I'm new to the stack Debugging
   Explains an error in plain English: what it means, three common causes, how to tell which one fits, and the usual fix.
4. Review AI-generated code Code review
   Reviews code an AI wrote for invented APIs, swallowed errors, weak tests and similar problems, then rates it merge, merge after fixes, or rewrite.
5. Coding agent system prompt (safe defaults) Agent system prompts & workflows
   Standing rules for a coding agent: read the repo first, keep changes small, run your checks, and stop to ask if the task is ambiguous or too large.
6. Fact-check this claim Research
   Splits a claim into checkable parts and marks each one true, false, misleading or unverifiable. The model notes recommend web browsing.
7. Edit for clarity, keep my voice Writing
   Edits a draft for clarity and grammar, keeps your wording where it works, does not add new ideas, and lists the main changes.
8. Landing page copy for an indie product Marketing for indie makers
   Drafts a headline, benefits, how-it-works, FAQ and a call to action from your product, audience, outcome and price.
9. Small-model friendly summariser Local LLM (small models)
   Asks for exactly five short bullets, using only facts from the text you paste. The model notes say it is tuned for small local models.
10. YouTube video script for a game Gaming & content creation
    Writes a timed script with a hook, segments and on-screen notes, and marks game facts you should double-check.

Browse the full library →

Collections by situation

The same free prompts, grouped for students, small-business admin, writers, job hunting, and everyday tasks. Each page links to prompts that are already in the library.

• For students. Reading, claim checks, draft edits, and plain-English explanations of errors and code.
• For small-business admin. Inbox triage, follow-ups, status notes, review replies, and document comparisons.
• For writers. Edits, posts from your notes, docs, a content brief, and short emails.
• For job hunting. Edit a letter, send a short note, follow up, and compare two versions of a document.
• For everyday life. Email, shortening, summaries, claim checks, and redacting personal details.

All collections →

Every prompt you can open on this site is free. The Pro Pack is a separate one-time download (US$12): 209 prompts as Markdown and JSON, 10 agent workflows, and 12 slash commands for Cursor and Claude Code.