A system prompt is the instruction text supplied to a language model before any user input, usually by the application developer rather than the end user. It sets the model's role, tone, output format, allowed topics and rules, and it is sent with every request in a conversation.
Why it matters
The system prompt is the main way to shape a general model into a specific product: a support bot, a code reviewer, a translator. Models usually give it more weight than ordinary user text, but that is a tendency rather than a guarantee. It counts toward the context window and your token bill on every call, so long prompts have a real cost.
Example
System: You are a support assistant for an uptime monitoring service.
Answer only questions about monitoring. Keep replies under 120 words.
If asked about billing, direct the user to the billing page.
User: Why did my check fail?Common mistakes
- Putting secrets such as API keys in it. Users can often coax a model into revealing its system prompt.
- Relying on it as a security boundary. Prompt injection can override instructions, so enforce permissions in code.
- Writing long, contradictory rule lists that the model follows inconsistently.
Related
Get notified when this term updates
Plus occasional practical tips on uptime and monitoring, nothing else. Unsubscribe anytime.
This page is general information written by our editorial team and updated when documentation changes. Servers and providers differ, so test changes safely first. See our editorial policy and disclaimer.