How sujeetgund/laptopwise works
Project type and purpose
• Type: Full-stack, AI-powered web application.
• Purpose: Allow end-users to enter laptop specs via a Next.js/React UI, validate them with an AI flow (Genkit + LLM), then call an external ML-model API to estimate market price.
File structure
• Top-level:
• package.json / tsconfig.json / next.config.ts / tailwind.config.ts → Node/Next.js configuration
• public/ → Static assets (images, fonts)
• src/app/ → Next.js App Router pages and root layout
• src/components/ → UI building blocks, ShadCN/Tailwind based components, section-level and layout-level components
• src/ai/ → Genkit flows (validate specs, predict price) + Genkit initialization
• src/lib/ → Constants & utility functions
• src/hooks/ → Custom React hooks
• No internal database, no separate “backend” folder—AI-flow server (Genkit) functions as the business-logic / microservice layer.
Insights from README
• Confirmed tech stack: Next.js (App Router), React, TypeScript, ShadCN UI, Tailwind, Genkit (LLM flows), external Python ML model on a cloud endpoint, Google AI credentials for Genkit.
• Workflow:
1. User fills form (react-hook-form + zod)
2. Call validate-laptop-specs flow (Genkit + LLM)
3. If valid, call predict-laptop-price-flow (Genkit → external ML API)
4. Result + insights