Examples — real raw → enhanced pairs
Each example shows the same idea twice: the rough version you might actually type, and the structured prompt Pixiy would produce for Claude, ChatGPT or Gemini. Click any card to see the full enhanced output plus the improvements that mattered.
- Code
- Marketing
- Writing
- Debug
- Research
- Education
- Career
- Product
- CodeClaude · XML
Code review request for Claude
You want Claude to review a pull request and return findings you can actually triage — not a wall of prose.
View example → - MarketingGPT · numbered
Product launch email for GPT
You need a launch email to convert Free users to Pro — and you want GPT to nail it without three rounds of editing.
View example → - WritingClaude · XML
Blog post outline for Claude
You want a useful long-form blog post outline, not a generic table of contents.
View example → - DebugClaude · XML
Bug debugging help from Claude
You hit a confusing bug and want Claude to debug with you — not just guess at the symptom.
View example → - ResearchGemini · long-ctx
Research brief for Gemini long-context
You need a balanced research brief Gemini can answer thoroughly — using its long context strength.
View example → - EducationGPT · numbered
Lesson plan for GPT
A teacher wants GPT to draft a one-hour lesson plan that maps to real learning outcomes.
View example → - ProductGemini · long-ctx
Product description for Gemini
You have rough product notes and want Gemini to draft a store-ready description.
View example → - CareerClaude · XML
Cover letter draft from Claude
You're applying for a role and don't want a stock cover letter the recruiter will skim past in 3 seconds.
View example →