We test prompts across 80 AI models — ChatGPT, Claude, Midjourney, Sora, Suno, and more — so you don't have to. Edits, votes, and version history. Always public.
760
prompts
80
AI models
52
categories
5
domains
Text, code, video, audio — not just pretty pictures. Browse top-rated prompts from each of the five domains, tested across the models that matter.
Rewrite the following email to be professional, clear, and effective. Follow these rules: **Tone**: [Choose: friendly-professional / neutral-professional / formal] **Guidelines**: 1. Open with a clear purpose statement — the reader should know why they're reading within the first sentence 2. Use short paragraphs (2-3 sentences max) and bullet points for multiple items 3. Remove filler words, hedging language ("I was just wondering if maybe..."), and unnecessary apologies 4. Replace vague requests with specific, actionable asks including deadlines 5. Close with a clear next step or call to action 6. Keep the total length under 150 words unless the content requires more 7. Preserve the sender's original intent and any critical details — do not add information that wasn't in the original **Format your response as**: - **Subject line**: (improved) - **Body**: (rewritten email) - **Changes made**: (brief list of what you changed and why) **Original email to rewrite**: [paste email here]

Each domain has its own quirks. We track which models work, where they fail, and what to write to get a good result on the first try.
Highest-voted prompts right now — copied, forked, and refined the most.
Upload a photo, pick a style — Ghibli, Pixar, action figure, synthwave, anime. Same prompt, your face.
Dynamic anime-style magical transformation with particle effects, costume change, and dramatic camera work.
Transform a screenshot into a polished App Store/Play Store listing image. Device mockup, gradient background, feature callouts — ready to publish.
Other sites give you a prompt and wish you luck. We give you a prompt that's been tested, edited, and improved by people who actually use it.
Every prompt gets better over time. Edits, votes, and peer review surface the version that actually works.
Full revision history with diffs. See exactly what changed, why, and by whom. Nothing is ever lost.
Practitioner-written guides on prompt design, plus per-model field notes. Read these and stop guessing.
A beginner-friendly guide to writing effective prompts for AI models. Learn the fundamentals that work across ChatGPT, Claude, Midjourney, and more.
How to use chain-of-thought (CoT) prompting to dramatically improve reasoning in ChatGPT, Claude, and other language models.
Learn how negative prompts work in Stable Diffusion, SDXL, and other image models. Includes templates and common patterns.
Multi-step prompt pipelines — one prompt feeds the next. Research → outline → draft → polish, without juggling tabs.
Promptibus MCP gives your AI agent real knowledge about 80 generative models — syntax, parameters, pricing, and routing. One install, every model.
recommend_modelBest model for a task, budget, and quality target
format_promptModel-native syntax with correct parameters and tags
compare_modelsSide-by-side on speed, cost, quality, and features
guidance_scale: 3.5Install the Promptibus Chrome extension and your bookmarked prompts live in the toolbar. Click the icon on any tab, hit Copy, paste into ChatGPT, Claude, Midjourney — anywhere. No context switching, no tab hunting.
Chrome Web Store listing coming soon. Until then: download ZIP → unzip → chrome://extensions → enable Developer mode → Load unpacked.

Every prompt is visible in full — no paywall on content. Ever. Paid tiers add unlimited copies, on-site runs, and creator analytics.
Unlimited copies. No ads. 10 credits/mo to run prompts on site.
See full comparison · Cancel anytime · No credit card for free account
















You are a senior debugging specialist. When given a bug report, you follow a rigorous diagnostic process — you never guess. **Input required**: - Bug description (what happens vs. what should happen) - Relevant code snippet(s) - Error message or stack trace (if any) - Steps to reproduce **Your debugging process**: 1. **Reproduce mentally**: Walk through the code path step-by-step with the given inputs. State what each line does. 2. **Isolate**: Identify the exact line or interaction where expected behavior diverges from actual behavior. 3. **Hypothesize**: List 2-3 possible root causes, ranked by likelihood. For each: - Why it could cause the symptom - How to verify or rule it out 4. **Fix**: Provide the minimal fix for the most likely cause. Explain: - What the fix changes - Why it works - What edge cases to test 5. **Prevent**: Suggest one thing (test, assertion, type, lint rule) that would have caught this bug earlier. **Rules**: - Never say "try this" without explaining why - If you need more information, ask specific questions before guessing - Always consider: race conditions, null/undefined, off-by-one, type coercion, stale closures, and encoding issues
You are a senior debugging specialist. When given a bug report, you follow a rigorous diagnostic process — you never guess. **Input required**: - Bug description (what happens vs. what should happen) - Relevant code snippet(s) - Error message or stack trace (if any) - Steps to reproduce **Your debugging process**: 1. **Reproduce mentally**: Walk through the code path step-by-step with the given inputs. State what each line does. 2. **Isolate**: Identify the exact line or interaction where expected behavior diverges from actual behavior. 3. **Hypothesize**: List 2-3 possible root causes, ranked by likelihood. For each: - Why it could cause the symptom - How to verify or rule it out 4. **Fix**: Provide the minimal fix for the most likely cause. Explain: - What the fix changes - Why it works - What edge cases to test 5. **Prevent**: Suggest one thing (test, assertion, type, lint rule) that would have caught this bug earlier. **Rules**: - Never say "try this" without explaining why - If you need more information, ask specific questions before guessing - Always consider: race conditions, null/undefined, off-by-one, type coercion, stale closures, and encoding issues
ChatGPT, Claude, Gemini
BrowseTurn your pet into a collectible action figure! Complete with toy packaging, pet-specific accessories, and an adorable character bio.
One prompt, tested everywhere. Compare results across Midjourney, Flux, DALL-E, Stable Diffusion, and more.
Real outputs from real users. Every prompt comes with a gallery of actual results, tagged by model and settings.
Prompts organized by what you actually do — not just by which model spits them out.
Write copy that converts, plan content that ranks, run campaigns that scale
Ship cleaner code, squash bugs faster, automate the boring parts
Go from concept to visual in seconds — product shots, characters, environments
Scripts, hooks, outlines, music — fuel your content pipeline across every platform
Build lesson plans, quizzes, and explainers that actually engage students
From pitch deck to financial model — move fast with AI-assisted strategy
Research → Outline → Draft → Edit & Polish
The best prompts come from people who use them daily. Share yours — or make an existing one better.