System Prompt Generator

A system prompt is the standing instruction a chatbot reads before every message: who it is, what it is for, how it talks, and where its limits are. It is what turns a general model into a support agent, a study tutor, or a brand voice. Writing one from scratch is fiddly because the useful parts are easy to forget until the assistant misbehaves.

This generator asks for the assistant's name, purpose, audience, personality, the things it must always do, the things it must never do, a response format, and one example exchange. It assembles those into a clean system prompt you can paste into custom instructions, a GPT builder, an API call, or a ChatUp chat. It runs in your browser and nothing you type leaves your device.

How to use it

1. Name the assistant and state its purpose in one or two sentences.
2. Describe the audience and the personality or tone you want it to keep.
3. List what it must always do and what it must never do, then pick a response format.
4. Add one example exchange, copy the result, and paste it wherever your assistant lives.

What belongs in a system prompt

A system prompt sets defaults the user should not have to repeat. The strongest ones are short on adjectives and long on specifics: what the assistant does, who it serves, how it should answer, and the handful of rules that matter most. Vague instructions like "be helpful" add nothing, because the model already tries to be.

The must-never list deserves the most thought. Models follow clear prohibitions well, so name the exact behaviors you want to rule out: inventing facts, giving medical advice, quoting prices, switching languages, or answering questions outside the assistant's job.

• Identity: a name and a one-line job description
• Audience: who is talking to it and what they already know
• Voice: two or three concrete adjectives, plus an example of how it opens
• Always: the steps it should take on every reply, such as asking a clarifying question first
• Never: the specific behaviors that are off limits
• Format: length, structure, and whether to use headings, lists, or code blocks
• Example: one short exchange that shows the standard

System prompt vs. regular prompt

A regular prompt is a single request; a system prompt is the frame every request is read through. In most chat apps the system prompt lives in settings under names like custom instructions, personality, or instructions for the GPT. Through an API it is a separate field sent with each conversation.

Because it applies to everything, a system prompt should avoid task-specific detail. Put the persona, rules, and format there, and leave the actual questions to the chat. If you find yourself rewriting the system prompt for each task, the AI Prompt Generator is the better tool.