Project

FamilyCookbook

A recipe application with semantic search and AI image generation. Angular, .NET 8, PostgreSQL with pgvector. In production and in daily use.

Something I've always wanted to build. AppFactory made it simple and quick, and me and my wife use it all the time.

Semantic Search — live demo

familycookbook.home / recipes

Walker Family Cookbook

Search by name, ingredient, or describe what you feel like… Ctrl+K

Quiche Lorraine

45 min

FrenchBaking

Geoff8.5

Helen9.0

Fish and Chips

50 min

BritishComfort

Geoff9.5

Helen8.0

Lemon Drizzle Cake

55 min

BakingDessert

Geoff8.0

Helen9.5

Spaghetti Carbonara

30 min

ItalianPasta

Geoff9.0

Helen8.5

Chicken Tikka Masala

50 min

IndianComfort

Geoff9.0

Helen7.5

Mushroom Risotto

35 min

ItalianVegetarian

Geoff8.0

Helen—

AI Image Generation

Every recipe, beautifully presented

Recipes without images — or with rough photos — can be enhanced with AI generation. These are real generated images from the app. Drag to compare.

Uploaded photo AI Generated

Quiche Lorraine — original uploaded photo vs AI-generated result. Real images from FamilyCookbook. Drag to compare.

Features

Built for real use

Meal Planner & Shopping List

Plan the week. The app generates a consolidated shopping list across all planned meals — deduped, sorted, ready to use.

Cook History & Versioning

Every cook is recorded. Variations accumulate as a version history — promote any version to become the canonical recipe.

The Geoff Filter

A suggestion queue weighted by personal preference and how recently each recipe was cooked. Solves "what should I make?" without randomness.

Ratings & Ingredient Scaling

Per-cook ratings tied to the specific version made. Servings adjust in real time across all stages — built into the cook flow.

Built with

Angular TypeScript .NET 8 C# PostgreSQL pgvector OpenAI Embeddings Docker GitHub Actions

github.com/Geoff-Walker/FamilyCookbook