Optimization Modes
The Prompt Optimizer offers five distinct modes, each applying a different strategy to improve your prompt.
Enhance & Clarify
The default mode. Improves clarity, specificity, and structure while preserving your original intent.
Best for: General-purpose prompt improvement, making vague prompts more specific.
What it does:
- Adds specificity where the prompt is vague
- Improves instruction structure for better AI understanding
- Preserves
{{variables}}exactly as they appear
Structured Format
Reformats your prompt into a clear, structured layout using sections:
- Role — Who the AI should act as
- Context — Background information
- Task — What to accomplish
- Format — Expected output format
- Constraints — Rules and limitations
Best for: Complex prompts that need clear organization, system prompts.
Chain of Thought
Adds step-by-step reasoning instructions to help the AI think through problems methodically.
Best for: Logic-heavy tasks, math problems, multi-step analysis, debugging.
What it does:
- Breaks down the task into logical reasoning steps
- Adds explicit "think step by step" instructions
- Structures the reasoning flow
Few-Shot Examples
Adds 2-3 example input/output patterns that demonstrate the expected behavior.
Best for: Tasks where showing examples is more effective than describing rules — formatting, classification, extraction.
What it does:
- Adds realistic input/output examples
- Covers typical cases and edge cases
- Uses clear "Example Input:" / "Example Output:" labels
Model-Specific
Optimizes for a specific AI model's strengths and conventions.
Target models:
- Claude — Uses XML tags, direct instructions, system prompt patterns
- GPT — Uses clear delimiters, role-play setup, explicit format instructions
- Gemini — Uses structured formatting, clear task decomposition
Best for: When you know which AI model will receive the prompt and want to maximize its effectiveness.