What-to-eat is an AI-powered recipe generation platform built with Vue that creates dishes from eight major Chinese cuisines plus international cooking styles.
The platform solves the problem of meal planning and recipe discovery by leveraging AI to generate professional recipes based on ingredients and cuisine preferences. It generates complete recipes with detailed cooking instructions, provides nutritional analysis with health scores, creates dish images via AI, recommends wine pairings, and designs custom sauce formulas. The tool includes a recipe collection system for saving favorites and a playful fortune-telling feature based on dietary preferences. A dynamic configuration system allows users to switch between different AI service providers compatible with the OpenAI standard, adjust model parameters like temperature and timeout settings, and test API connections without restarting the application.
The platform suits developers and home cooks seeking comprehensive cooking guidance with AI-generated content. It works best for users who want professional-quality recipes with nutritional insights and visual presentation, and who may want to experiment with different AI models to find preferred recipe styles and creativity levels. The tool emphasizes that recipe quality and professional detail vary significantly across different AI models, recommending high-quality models for better results.
The project was developed using vibe coding methodology through the Kiro editor, implementing the full development workflow from requirements analysis through architecture design to code implementation. The codebase is built on Vue 3.4 with TypeScript, styled with Tailwind CSS, and uses Vite as the build tool. Deployment targets Vercel and Netlify platforms. The project maintains separation between recipe generation and image generation model configurations, with user settings persisted locally and validated through built-in API testing functionality.
