AI Prompt Checker

Paste your AI prompt below and get an instant quality score with actionable suggestions to make it better.

100% private — runs in your browser No sign-up required Instant analysis

Your Prompt

0 characters

What makes a great AI prompt?

Set a role. "You are a senior copywriter..."
Be specific. Avoid vague words like "good" or "nice".
Specify format. List, table, JSON, word count.
Define the audience. Who's the output for?
Add constraints. Tone, length, style.
Give examples. Few-shot examples dramatically improve quality.

Ready when you are

Paste your prompt on the left and click Analyze to get an instant quality score and improvement tips.

0 /100

—

Analysis pending

Score Breakdown

Unlock your fully rewritten prompt

You've got your score and the issues above. Enter your email and we'll instantly reveal a fully rewritten, higher-scoring version of your prompt — plus a one-page prompt cheat-sheet, sent to your inbox.

Your rewritten prompt

Paste this into ChatGPT, Claude, or Gemini. Replace any [bracketed] placeholders with your specifics.

Questions answered

Frequently asked questions

Everything you might be wondering about the AI Prompt Checker.

How does the prompt scoring work?

Each prompt is checked against 8 heuristics — length, role definition, specificity, output format, audience, constraints, examples, and clear directive. Each one is weighted by impact, and the total is graded as Excellent (85+), Good, Decent, Needs Work, or Poor.

Will this work for ChatGPT, Claude, and Gemini?

Yes. The heuristics measure universal prompt-quality principles that apply to every major LLM. A prompt that scores 90+ on our checker will produce strong output on any of them.

Why is my "role definition" marked as missing?

Starting a prompt with "You are a [role]..." dramatically improves output quality because it anchors the model's style, tone, and depth of knowledge. The checker looks for patterns like "You are", "Act as", "Imagine you are" — if none are present, role definition scores zero.

Are my prompts stored or shared anywhere?

No. Analysis runs entirely in your browser using JavaScript. Your prompts never leave your device — we don't have a server-side log, database, or analytics that captures the prompt text.

How do I get a 100/100 score?

Hit all 8 checks: set a role, be specific (use numbers and proper nouns), specify the output format (word count, structure), define the audience, add constraints (tone, things to avoid), include 1–2 examples, and start with a clear action verb (Write, Create, Generate).

See it in action

A worked example

Real input, real output — so you know what to expect before you run it yourself.

Vague prompt vs. a 95+ score prompt

Sample input

tell me about AI

Sample output

Score: 12/100 (Poor). Suggestions: Add a role ("You are an AI researcher..."), specify the output format (word count, paragraphs vs list), define the audience, and start with an action verb. Try: "You are an AI researcher. Write a 400-word explainer about how large language models work, for non-technical marketing managers. Use 3 short paragraphs and one analogy."

More free tools