Mastering Prompt Fundamentals
A well-structured prompt is the key to unlocking an LLM's potential. It's an iterative process of writing, testing, and refining. Below are the key components of an effective prompt.
Role & Objective
Define the LLM's persona and its overall goal. (e.g., "You are a helpful culinary assistant...")
Instructions & Rules
Provide clear, specific directives on what to do and what to avoid. Use lists for clarity.
Context
Supply the background information, data, or text the LLM needs to perform the task.
Examples (Few-Shot)
Offer input-output pairs to guide the model on format, style, and detail.
Reasoning Steps
Instruct the model to "think step by step" to break down complex problems before answering.
Output Formatting
Explicitly define the desired structure of the response, often using Markdown.
Anatomy of a Prompt
While every prompt is different, this gives a general idea of component weight in a complex system prompt.
The Prompt Crafter
First, generate a new scenario. Then, fill out the fields below to build your system prompt. Use the ✨ buttons to get guiding questions from your AI coach.
1. Start Your Practice
Click the button below to get a new, random scenario to build a prompt for.
Describe your own AI assistant scenario to practice with.
The Query Tester
Once you've drafted your prompt, test its performance. Add diverse queries to the list below and run them against the Gemini API.