Roast My Prompt
Paste any ChatGPT, Claude, or AI prompt and get an instant quality score across five dimensions — with specific fixes and an improved version ready to copy. Free prompt checker, no sign-up.
What gets scored
Clarity — does the task verb appear upfront, or is it buried in a soft opener?
Specificity — are requirements concrete, or does it rely on vague words the model has to interpret?
Context — is there a role frame, target audience, or background that helps the model calibrate?
Format — does it specify how the output should be structured and how long?
Constraints — does it set tone, and specify what to avoid or exclude?
Why prompts fail
Most AI prompts produce mediocre output not because the model is incapable, but because the prompt leaves too many decisions to the model. When you don't specify format, the model picks one. When you don't set context, the model assumes the most generic interpretation. When you don't add constraints, it hedges and adds filler.
The five dimensions this tool scores are the same ones used by the ChatGPT prompt generator and Claude prompt generator to build better prompts from scratch.
Frequently Asked Questions
How does this prompt quality checker work?
It analyses your prompt across five dimensions — clarity, specificity, context, format, and constraints — and scores each one based on what's present and what's missing. The improved version applies fixes automatically.
Does it work for ChatGPT, Claude, and other models?
Yes — the five dimensions apply to any large language model. The principles that make a prompt work (clear task, specific requirements, format instructions) are consistent across ChatGPT, Claude, Gemini, and others.
What score should I aim for?
7 or above consistently produces noticeably better output. Scores of 4–6 usually mean one or two specific dimensions are dragging down otherwise solid prompts — the fixes section will point to exactly what to change.
Can I rate my prompt more than once?
Yes — paste in your improved version and roast it again. The score should climb as you apply the fixes. It's a useful iteration loop for prompts you use repeatedly.