Culinode

Culinode started from a simple problem: recipes scattered across bookmarks, screenshots, messaging apps, and paper notes. I wanted one place to store them, plan meals around them, and auto-generate shopping lists — self-hosted, so the data stays mine. I wrote a PRD in Claude Code defining the core entities (recipes, ingredients, meal plans, shopping lists) and the monorepo structure (React frontend, Express backend, shared TypeScript types). The first working version — recipe CRUD with URL import — was running in Docker within the first week.

The meal planning system was the biggest design challenge. A weekly calendar needs to show enough information per meal (recipe name, servings, cook time) without becoming overwhelming. The solution was a day-by-day vertical layout with color-coded meal type badges (lunch, dinner) and inline stat cards that aggregate nutrition and prep time across the full plan. Shopping lists generate automatically from any meal plan, grouping ingredients by category (Protein, Dairy, Produce, Pantry) and consolidating duplicates — if two recipes both need garlic, you see one line item with the combined quantity.

The multi-provider AI integration mirrors the same pattern I used in Cortex: let users pick their preferred provider (OpenAI, Anthropic, Google AI, or local Ollama) and configure it per-account. AI features are entirely optional — the app works fully without any API key. The gamification layer (levels, points, streaks, 16 achievements) was added to make the app stickier for daily use — cooking habits form faster when there’s a visible progress loop.