Paste any prompt. Get an instant 1 to 10 score, a list of weaknesses, and a rewritten version you can drop straight into Claude, ChatGPT, or your own agent.
Weekly AI dev tutorials, prompt patterns, and tool reviews. Free.
Estimate token counts and cost before you ship a prompt to production.
Cut your Claude API bill by 80 to 90 percent with prompt caching. Full setup walkthrough.
It reads your prompt, points out weaknesses like vague language, missing constraints, or missing examples, gives you a score from 1 to 10, and writes a stronger drop-in replacement.
Anthropic's Claude. The response streams directly from the Claude API to your browser.
No. We do not log, save, or train on submitted prompts. The text is sent to Claude for the single critique and discarded.
Yes. To keep the tool free, each IP can run 5 critiques per 15 minutes. Sign up for the newsletter if you want a heads-up when we ship a higher-volume version.
Clear role, specific task, explicit constraints (length, tone, format), success criteria, and one or two examples when the output is non-trivial. The critique pinpoints which of these your prompt is missing.