REAL COMPARISONS
Prompt Optimization Examples
See how PromptCraftin refines, structures, and compiles simple conversational inputs into professional, high-fidelity prompt templates. Compare the vague before prompts with our compiled after prompts across different scenarios.
1. Copywriting & Content Marketing
❌ Vague Input (Before)
"write a sales email for a new water bottle"
The Problem: Lacks target audience context, brand tone guidelines, visual formatting instructions, and conversion constraints. Resulting output will be generic and uninspired.
✅ Compiled Prompt (After)
Role: Expert Direct-Response Copywriter.
Task: Write a high-converting cold sales email promoting our new eco-friendly, double-walled vacuum-insulated stainless steel water bottle.
Context: Target audience is busy professionals who value sustainability and hot/cold longevity. The bottle keeps drinks cold for 24 hours.
Constraints:
- Tone must be energetic yet professional.
- Subject line must be under 8 words.
- Keep the body text under 150 words.
- Include exactly one bold Call-to-Action link at the end.
- Avoid clickbait and exaggerated claims.
Output Format: Plain text formatted with Subject, Body, and Call-to-Action placeholders.
The Benefit: Clear roles and strict formatting rules force the AI model to output highly professional, conversion-focused, and brand-aligned content.
2. Software Engineering & Coding
❌ Vague Input (Before)
"write a react hook for localstorage"
The Problem: No specification of state syncing, TypeScript interface support, window-undefined server-side rendering (SSR) safety, or error handling. Will produce bare-bones, fragile code.
✅ Compiled Prompt (After)
Role: Senior React Developer.
Task: Create a custom React hook `useLocalStorage` in TypeScript.
Context: The hook must synchronize state with window.localStorage. It must support state updates, JSON serialization, and gracefully handle server-side rendering environments where the window object is undefined.
Constraints:
- Write code matching clean functional programming styles.
- Include explicit error-catching blocks for disabled localStorage settings.
- Implement state syncing across multiple tabs.
- Must provide strict TypeScript type-hints.
Output Format: Return a single, copyable TypeScript code block with explanatory comments and usage examples.
The Benefit: Forces the AI to account for complex real-world edge cases (like SSR and disabled storage) that standard generation ignores.
3. Midjourney & AI Image Generation
❌ Vague Input (Before)
"draw a cool glass building in the forest"
The Problem: Lacks detail regarding cinematic lighting, camera lenses, visual style, color palettes, and rendering engine flags (e.g., Midjourney `--v 6.0`).
✅ Compiled Prompt (After)
Core Concept: An architectural glass forest villa.
Style: Biophilic architecture, modern structuralism, cinematic rendering.
Lighting: Soft twilight golden-hour rays, dramatic long shadows, interior warm glowing light.
Composition: Eye-level wide cinematic landscape view, wet mossy forest floor reflection, pine trees framing.
Camera Parameters: Shot on Hasselblad 500c, 35mm lens, f/1.8 aperture, extreme hyper-detail, natural color grading.
Aspect Ratio / Midjourney Flags: --ar 16:9 --v 6.0 --style raw --stylize 250
The Benefit: Calibrates lighting, aspect ratio, styles, and rendering engine specific weights to get flawless, breathtaking photorealism instead of generic shapes.