Cheat Codes for better design and code
Copy-ready ChatGPT prompts, design recipes, and engineering workflows for turning an idea into a polished, working product.
Create your exact prompt
Combine a product goal, visual direction, and implementation stack to generate one of 2,560 detailed ChatGPT design codes.
Act as a senior product designer and frontend engineer. Design and implement a SaaS landing page using React with TypeScript and Tailwind CSS. Visual direction: editorial and typography-led. Before writing code: 1. Define the target user, primary task, and information hierarchy. 2. Describe the design concept in 3 concise principles. 3. List the reusable components and responsive behavior. Implementation requirements: - Use purposeful typography, spacing, color, and hierarchy. - Include polished loading, empty, error, and success states where relevant. - Meet WCAG AA contrast and keyboard-navigation expectations. - Make the layout excellent on mobile, tablet, and desktop. - Use realistic content instead of generic placeholder copy. - Follow existing project patterns and avoid unnecessary dependencies. Deliver the complete implementation, explain the important design decisions, and provide a short verification checklist.
Find your next shortcut
Start with a template, replace the bracketed context, and ask ChatGPT to work in small, verifiable steps.
Create a complete design system
Turn a visual direction into reusable tokens, components, and rules.
Act as a senior product designer and design systems engineer. Create a design system for [product] with: - A clear visual concept and 3 design principles - Color tokens with accessible contrast - Type scale, spacing scale, radius, and shadow tokens - Button, input, card, modal, navigation, and empty-state patterns - Responsive behavior for mobile, tablet, and desktop Return the system as implementation-ready CSS variables and component guidance. Keep the interface polished, cohesive, and easy to extend.
Build a high-converting landing page
Generate a page structure that communicates value before decoration.
Design a responsive landing page for [product] aimed at [audience]. Include: 1. A specific hero headline and supporting value proposition 2. One primary call to action and one low-friction secondary action 3. Social proof, benefits, product workflow, FAQ, and final CTA 4. Mobile-first layout, accessible headings, and visible focus states Use [framework] and [styling system]. Avoid generic gradients and filler copy. Explain the layout hierarchy before writing the code.
Design a useful analytics dashboard
Prompt for hierarchy, data density, responsive states, and meaningful empty states.
Create an analytics dashboard for [user role] tracking [metrics]. The dashboard must: - Surface the 3 most important decisions at the top - Use cards only when they improve scanning - Include loading, empty, error, and no-permission states - Support keyboard navigation and readable color contrast - Adapt the table and charts for narrow screens Use realistic sample data and explain why each visualization fits its metric. Build the UI with [React/Tailwind/etc.] and keep components reusable.
Ship an accessible form
Create a form with clear validation, keyboard support, and useful feedback.
Build a production-ready [form name] in React and TypeScript. Requirements: - Use labels connected to every field - Validate with [Zod/your validation library] - Show inline errors without relying on color alone - Preserve entered values after validation errors - Disable submit while saving and prevent duplicate submissions - Announce success and server errors accessibly - Include loading, empty, and network-failure behavior Match the existing component patterns in [reference file] and add focused tests.
Make a card responsive
A reliable prompt for turning a desktop card into a strong mobile component.
Refactor this card component for responsive behavior. Goals: - Preserve the information hierarchy at every width - Keep the primary action visible without awkward wrapping - Use a stacked layout below [breakpoint] - Handle long titles, missing images, and empty metadata - Keep touch targets at least 44px - Match the existing spacing, typography, and interaction patterns First explain the layout changes, then update only the necessary component files.
Debug without guessing
Force a clean investigation with evidence, hypotheses, and a small fix.
Investigate this bug systematically. Expected behavior: [describe it] Actual behavior: [describe it] Exact error or logs: [paste logs] Relevant code: [paste the smallest useful section] Do not rewrite the feature. First: 1. List the likely causes in order 2. State what evidence would confirm each cause 3. Identify the smallest safe fix 4. Suggest a regression test and verification command
Review code for security risks
Use this before shipping authentication, payments, forms, or user-generated content.
Review this change as a security-focused senior engineer. Check for: - Authentication and authorization gaps - Injection, XSS, CSRF, and unsafe redirects - Secrets or personal data exposed in logs or responses - Missing server-side validation and rate limits - Insecure file uploads, dependencies, or error messages - Race conditions and unsafe assumptions at trust boundaries For every finding, include severity, evidence, exploit scenario, and a minimal fix. If something is safe, explain what control makes it safe.
Generate a typed API endpoint
Keep request validation, business logic, and response errors explicit.
Add a [HTTP method] endpoint for [resource] to this [framework] API. Context: - Authentication: [method] - Database: [database/ORM] - Existing route pattern: [file or example] - Response format: [shape] Include: - Request schema validation - Authorization checks - Explicit success and error responses - Database failure handling - Types shared between client and server where appropriate - A test for success, invalid input, unauthorized access, and missing data Follow existing project conventions and do not introduce a new library.
Model data before coding
Get a clear schema and migration plan before building the UI.
Design the database model for [feature]. Users need to [user workflows]. Identify: - Entities, fields, types, defaults, and constraints - Relationships and delete behavior - Indexes for the expected queries - Uniqueness and ownership rules - Audit fields and privacy-sensitive data Return: 1. A concise data model 2. The migration/schema code for [ORM] 3. Example queries for the main reads and writes 4. Risks, tradeoffs, and a rollback plan
Make a page search-ready
Create useful metadata and content structure without keyword stuffing.
Optimize [page] for people searching for [topic]. Create: - One clear title and meta description - A specific H1 and logical H2 structure - Helpful introductory copy that answers the user's intent - Internal-link suggestions - FAQ questions worth answering - Open Graph and social preview copy Keep the writing natural, accurate, and useful. Do not make unsupported claims or repeat keywords unnaturally.
Find the biggest performance wins
Prioritize measurable improvements instead of premature optimization.
Audit this feature for performance bottlenecks. Context: [page, user flow, traffic, and device constraints] Code or profiling data: [paste evidence] Separate findings into: - User-visible issues - Network and data-loading costs - Rendering and bundle costs - Database or server costs Rank each by impact and effort. Recommend the smallest high-impact changes, include a measurement plan, and avoid optimizing code without evidence.
Turn requirements into tests
Convert acceptance criteria and edge cases into a focused test plan.
Create a test plan for [feature] from these requirements: [paste requirements or acceptance criteria] Cover: - The main happy path - Validation and boundary values - Loading, empty, error, and retry states - Authentication and permission differences - Keyboard and responsive behavior - Regression risks from existing functionality Return a prioritized list of tests with setup, action, and expected result. Use the project's existing test framework and patterns.
How to get better results from every code
1. Add context
Name your stack, existing patterns, user, constraints, and the files that matter.
2. Ask for one slice
Small requests create changes that are easier to understand, test, and undo.
3. Verify the result
Review the code, run checks, and test edge cases before treating AI output as finished.