Prompt Generator

Transform raw tasks into structured AI prompts using 11 evidence-based frameworks including ROSE, CO-STAR, Chain of Thought, and Few-Shot prompting.

Refine your prompts,
elevate your work

Transform raw ideas into expertly-structured prompts. Work faster, think clearer, achieve better results.

No Login
No Subscription
Always Free
The Problem

The Old Way

  • Spend 10+ minutes trial-and-error testing
  • Get inconsistent results across frameworks
  • Struggle to follow framework best practices
  • Waste time on iterations instead of creation
  • Other tools lack framework support
The Solution

The Prompt Generator Way

  • Get polished prompts instantly
  • Access 12 proven frameworks
  • Guaranteed consistent output
  • Focus on ideas, not engineering
  • Only tool with framework options

12 Specialized AI Agents.
One Objective. Perfect Prompts.

These agents specialize in prompt engineering frameworks, each designed to refine your inputs into clear, structured prompts.

AUTO

ASSESS, DECIDE, REFINE

Best for: Not sure where to start. Let the AI choose for you.

ROSE

ROLE, OBJECTIVE, STYLE, EXEMPLAR

Best for: Tasks needing a clear persona and style example.

APE

ACTION, PURPOSE, EXPECTATION

Best for: Action + purpose + success criteria defined upfront.

RACE

ROLE, ACTION, CONTEXT, EXPECTATION

Best for: Grounded in realistic context and constraints.

RISE

ROLE, INPUT, STEPS, EXPECTATION

Best for: Structured multi-step, process-oriented tasks.

RTF

ROLE, TASK, FORMAT

Best for: Minimal overhead; output format is critical.

CO-STAR

CONTEXT, OBJECTIVE, STYLE, TONE, AUDIENCE, RESPONSE

Best for: Tone, audience, and communication matter.

COT

CHAIN OF THOUGHT

Best for: Transparent step-by-step thinking.

CRISPE

CAPACITY, ROLE, INSIGHT, STATEMENT, PERSONALITY, EXPERIMENT

Best for: Unlocking creative and novel ideas.

FEW-SHOT

EXAMPLES-BASED PROMPTING

Best for: Teaching by concrete example, not description.

REACT

REASONING + ACTING

Best for: Reasoning + action cycles; agent-style tasks.

HYBRID

MULTI-FRAMEWORK COMBINATION

Best for: High-stakes work; combines 2-4 frameworks intelligently.

Be as specific as possible for better results

Dictate

Generated Output

Your Refined Prompt

Generated system prompt will appear here...

Support

Your questions, answered

What is prompt engineering?

Prompt engineering is the practice of crafting clear, structured inputs to get better outputs from AI tools. A well-built prompt gives the AI a role, context, and constraints — reducing vague or off-target responses.

How is this different from just typing into ChatGPT directly?

Unstructured prompts produce inconsistent results. Frameworks add discipline: they ensure your prompt includes the right role, context, and output format. The result is more focused, useful AI responses — first time, every time.

Which framework should I choose?

It depends on your task. Click on any tile to learn more about what each framework does, or select AUTO if you're unsure — the AI will analyse your prompt and choose the best framework for you automatically. The best approach is to match your task type: use CO-STAR when tone and audience matter, Few-Shot when outputs need to match a specific style, CoT for logical reasoning, and Hybrid for complex multi-faceted problems.

Can I use the outputs for professional or published work?

Yes, with care. This tool generates prompt templates — what you do with the AI outputs is your responsibility. Always review AI responses critically and follow any relevant policies at your organisation.

Is my input stored or shared?

Your input is sent to an AI model to generate a refined prompt. We track only a simple counter of total uses — we do not log, store, or analyse the content of what you enter. Avoid entering sensitive or confidential information.

What if the output doesn't look right?

Refine your input. Be more specific about your goal, audience, or constraints, then regenerate. You can also switch frameworks — some work better for certain tasks. The Hybrid option is a good fallback if a single framework isn't working.

Help Us Improve

Your feedback shapes how we build. Share your thoughts in seconds.

We read every response. Thank you for helping us improve.

Framework