ChatGPT-Shortcut

Enables users to browse and filter a curated JSON-based prompt library across 13 languages (English, Chinese, Spanish, Arabic, Portuguese, etc.) using Docusaurus's built-in i18n system with client-side tag-based filtering. The system stores prompts as structured JSON objects with language-specific content, metadata, and category tags, allowing real-time filtering without backend queries. Filtering operates on prompt attributes like category, use-case, and difficulty level through React Context state management.

Allows users to create, edit, save, and organize custom prompts in a personal library using React Context API for state management and browser LocalStorage for persistence. Users can fork existing prompts from the catalog, modify them, and save them locally without backend infrastructure. The system maintains a User context that tracks favorites, custom prompts, and user preferences, with data persisted across browser sessions via LocalStorage.

Renders ChatGPT-Shortcut as a responsive web application using Ant Design 5.x components and custom React components, ensuring usability across desktop, tablet, and mobile devices. The Docusaurus framework handles responsive layout through CSS media queries and flexible grid systems, while Ant Design provides pre-built responsive components. The UI adapts to different screen sizes without requiring separate mobile or tablet versions.

Implements instant page loading through a custom Docusaurus plugin (plugins/instantpage.js) that preloads pages on hover or link focus, reducing perceived latency when navigating between prompts. The plugin likely uses the Instant.page library or similar approach to prefetch linked pages before the user clicks, creating a snappy navigation experience. Combined with Docusaurus's static site generation, this enables near-instant page transitions.

Enables users to share custom prompts with the community and contribute new prompts to the public catalog through a GitHub-based contribution workflow. The system uses a community-prompts page where users can view shared prompts, and contributions are managed via pull requests to the prompt.json file in the repository. The architecture leverages GitHub as the backend for version control, review, and merging of new prompts, with Docusaurus rendering the community content statically.

Provides browser extension and Tampermonkey userscript implementations that inject ChatGPT-Shortcut prompts directly into ChatGPT, Claude, and other LLM interfaces. The extensions use browser extension APIs to communicate with the main Docusaurus site, fetch prompts from the catalog, and inject them into the LLM chat interface via DOM manipulation. The userscript approach enables cross-browser compatibility without requiring formal extension store approval.

Defines and enforces a structured schema for prompts using TypeScript interfaces (LanguageData, prompt objects) that specify required fields like title, description, category, tags, and language-specific content. The system validates prompts against this schema during contribution and rendering, ensuring consistency across the catalog. Metadata includes multilingual content, difficulty levels, use-case categories, and contributor attribution, all stored in the prompt.json file with strict JSON structure.

Supports 13+ languages through Docusaurus's built-in i18n system combined with a custom JSON-splitting mechanism that separates language-specific prompt content. Each prompt stores language variants in a LanguageData structure, and Docusaurus automatically routes users to the appropriate language version based on browser locale or user selection. The system uses i18n configuration in docusaurus.config.js to define supported locales and default language, with translation resources organized in i18n/ directory structure.

Builds ChatGPT-Shortcut as a static site using Docusaurus 3.x, enabling deployment to multiple platforms (Vercel, Docker, Cloudflare Pages) with zero runtime dependencies. The build process uses Node.js and Yarn to compile React components, process i18n translations, and generate static HTML/CSS/JS artifacts. Deployment is automated through CI/CD pipelines that trigger on repository updates, with support for containerization via Docker for self-hosted deployments.

Provides feedback and comments functionality through a dedicated feedback page and comments component (src/pages/_components/Comments.tsx) that allows users to report issues, suggest improvements, or discuss prompts. The system integrates with GitHub Issues or a similar backend to track feedback, enabling community-driven improvements. Comments are rendered on individual prompt pages, creating a discussion thread around each prompt without requiring a custom backend.

Displays curated prompt collections and featured prompts through a showcase page (community-prompts.tsx) and ShowcaseCard components that highlight high-quality or trending prompts. The system uses metadata and community engagement metrics to surface the most useful prompts, creating a discovery experience beyond simple catalog browsing. Featured prompts are manually curated or algorithmically selected based on usage, ratings, or community feedback.

Tracks prompt usage, popularity, and performance metrics to identify trending prompts and inform curation decisions. The system likely collects analytics through browser events or server-side logging when users interact with prompts (copy, share, save, use). Metrics may include view counts, copy frequency, save counts, and community ratings, aggregated to surface high-performing prompts in the showcase and recommendations.