Term
System Prompt
A system prompt is the instruction that shapes a language model's behavior across an entire conversation — in contrast to the user prompt, which varies per message.
System Prompt — explained in more detail
A system prompt is the instruction given to a language model that shapes its behavior across an entire conversation — in contrast to the user prompt, which varies per message. Typical content includes the assistant’s role, style guidelines, available tools, response language and hard rules about what the model may or may not do. The system prompt is set by the application developer and stays active across every user turn.
Example / Practical context
A support bot might set in its system prompt: “You are the assistant for product X, you reply in English, you do not offer discounts and you refer legal questions to support.” Whatever the user asks, those constraints apply to every answer. Tools like Claude Code or Cursor likewise load an extensive system prompt that defines editor behavior (tool usage, coding style, safety rules).
Distinction from related concepts
What sets a system prompt apart from a user prompt is reach and visibility: it spans the whole conversation and is usually invisible to the end user. It also differs from fine-tuning because it does not modify the model — it is just context sent with every request, and it can be swapped out at any time.
Entdecke mehr
Effort level and deep thinking: two independent axes for AI tasks
Effort scales breadth, deep thinking scales depth. When each setting makes sense — with three clear examples and one rule of thumb.
GlossarChain-of-Thought
Chain-of-Thought (CoT) is a prompting technique that asks the model to spell out its reasoning in intermediate steps — boosting accuracy on multi-step tasks.
LexikonPrompt Security — Injection, Leaking, Guardrails
How prompt injection, prompt leaking, and jailbreaks work — and which defenses (guardrails, spotlighting, sanitization) actually help.