What is Prompt?

The instruction or input you give an AI model. Clear, specific prompts produce better, more reliable output.

A prompt is everything you send to an AI model in a request: the question, instructions, examples, and any background text you include. The model treats the whole thing as input and generates a continuation, so the way you phrase it directly shapes the answer. Vague prompts force the model to guess your intent; specific prompts about audience, format, length, and constraints get more useful output. In API usage, prompts are often split into a system prompt (standing instructions that set role and rules, usually written by the app developer) and a user prompt (the actual request). Many AI products you use daily are, underneath, a fixed system prompt wrapped around whatever you type. Prompts are counted in tokens, so longer prompts cost more and eat into the context window.

Example

"Write a product description" gets generic filler. "Write a 50-word product description for a $30 stainless steel water bottle, aimed at hikers, plain tone, no exclamation marks" gets something usable.

Why it matters

Two people using the same model can get very different results based on prompts alone, so prompt quality often matters more than switching tools. Browse the AI tools directory or the model leaderboard to put it into practice.