How flyingvaibhav/ai-course-generation works
Project type and purpose
• This is a Turborepo monorepo hosting:
• A full-stack Next.js application called “letskraack” under apps/letskraack
• Shared tooling packages under packages (UI component library, ESLint configs, TypeScript configs)
• The “letskraack” app provides AI-powered features (interviews, coding challenges, resume analysis, course management, discussion rooms), backed by Next.js API routes, Drizzle ORM, and external AI/voice services.
File structure
• Top-level:
• apps/letskraack: Next.js app containing:
– src/app/...: file-based routing for pages and API routes
– src/components/...: domain-specific React components
– src/services: client wrappers for AI, coding evaluation, TTS/STT, user services
– src/config, lib, constants, context, hooks: configuration, utilities, and state management
– drizzle.config.ts: DB connection (likely PostgreSQL) using Drizzle ORM
• packages/: shared packages:
– eslint-config: monorepo lint rules
– typescript-config: shared tsconfig JSONs
– ui: shared React UI primitives
• Root files (turbo.json, pnpm-workspace.yaml): Turborepo orchestration, workspace config
• Patterns:
• Monorepo with task orchestration (build/dev/test) via Turborepo
• Full-stack Next.js with file-based routing, serverless API routes
• Modular separation: UI library vs. app code vs. tooling configs
• Service-oriented: each feature area has its own API route + client service in code
README for additional insights
• Confirms monorepo layout: multiple apps/packages, uses TS, ESLint, Prettier, Turbo CLI commands, remote caching
Detailed capture of architectural elements
• Route groups (auth, workspace, ai-interview, coding, resume, discussion)
• API function endpoints (ai-chat, coding-check, resume-analysis, etc.)
• Service modules (aiModelServices, voiceToTextService, etc.)
• Data schema/config (drizzle.config.ts, schema.ts)
• Shared UI library as a single reusable block
• Pipeline steps: lint, type-check, build, test, cache