AI Prompt Optimizer

No. It is rule-based. It splits your prompt into sentences and sorts each one by keyword patterns: role statements such as 'act as', instructions that start with a verb, format requests, tone words, rules like 'don't' or 'must', and quoted examples. It then places them into the framework you choose. Your text is never sent anywhere.

No. It keeps your sentences as you wrote them (only capitalising the first letter and adding a full stop where one is missing) and reorders them under clear section labels. Sections it could not fill get an [Add …] placeholder so you can see exactly what is missing.

The score checks seven things good prompts usually include: a role, a clear task, context, an audience, an output format, constraints and an example. Each one found adds a point. A low score does not mean the prompt is bad, only that the AI will have to guess more.

Keyword rules can misread a sentence, for example a sentence that contains both a task and a tone. The output box is editable, so move or rewrite anything before you copy it. Each sentence's detected type is listed under the score so you can check.

RACE is a good default. CO-STAR suits writing where tone and audience matter, CREATE suits prompts with examples, SPARK suits problem solving, and APE or RTF suit short, simple requests.

The tool reads the first 20,000 characters of your prompt. That is far longer than most prompts.