How to use the Prompt Generator
1. Select a prompt template from the tabs — each is optimised for a different use case.
2. Fill in the fields. Leave them empty to use the placeholder example values.
3. Copy the generated prompt and paste it directly into your AI chat or API call.
What is it?
Prompt engineering is the practice of crafting inputs to language models to reliably produce high-quality outputs. Well-structured prompts dramatically improve accuracy, consistency, and controllability. This tool provides templates for six battle-tested prompting techniques used in production AI applications.
Chain-of-thought prompting improves reasoning for complex tasks. Role prompting leverages the model's domain knowledge. Few-shot examples guide output format. JSON output prompts ensure parseable structured responses. Summarization and code generation templates provide the right context and constraints to get professional-quality output every time.
Example
Task: What is the best database for a real-time chat app? Context: 100k DAU, horizontal scaling needed
Context: 100k daily active users... Question: What is the best database... Let's think through this step by step: 1. First, let me understand core requirements. 2. Then I'll consider key constraints and trade-offs. 3. Finally, I'll provide a reasoned recommendation.
Chain-of-thought prompts improve accuracy on complex reasoning tasks by 20-50% vs direct questions.
Benefits
Apply proven prompting techniques without memorising the patterns
Chain-of-thought, few-shot, role prompting, and ReAct templates are built in — select and customise rather than write from scratch.
Consistent prompt structure produces more predictable model output
Structured prompts with clearly defined roles, instructions, and examples outperform unstructured prompts across all major models.
Codify your best prompts as reusable templates
Document and standardise the prompt structure that works for your use case — easy to share, version, and iterate with your team.
Iterate on prompt structure systematically rather than randomly
Template-based prompts make it obvious which component to change when output quality drops — role, instruction, examples, or format.
Frequently asked questions
Which template should I use?
Chain-of-thought for complex reasoning and analysis. Role prompting for expert review/critique. Few-shot for classification and format-following tasks. JSON Output when you need parseable structured data. Summarizer for content distillation. Code Generator for programming tasks.
Why do empty fields use placeholder values?
The placeholders are working examples that produce a complete prompt instantly. You can see the structure and output immediately, then replace the placeholders with your actual content.
What is temperature and when should I change it?
Temperature (0-2) controls randomness. Use 0-0.2 for factual, deterministic tasks (JSON extraction, code generation). Use 0.7-1.0 for creative tasks (brainstorming, writing). Use 1.0+ for maximum creativity. Chain-of-thought prompts work best at 0.3-0.7.