AI Cookbook

Intermediate

ai creative frontend health

Click to zoom

AI Cookbook

Generate personalized recipes from your available ingredients with AI-powered recipe creation and beautiful food photography.

Transform your available ingredients into culinary masterpieces with the AI Cookbook. This intelligent recipe generator analyzes what you have on hand and creates complete, customized recipes tailored to your dietary preferences and cuisine choices.

Features at a Glance

Smart Ingredient Matching: Input what's in your kitchen and let AI do the magic
Beautiful AI Images: Every recipe comes with stunning, AI-generated food photography
Dietary Flexibility: Supports gluten-free, vegan, vegetarian, and more
Nutritional Insights: Get estimated calories, macros, and dietary information
Shopping Assistant: Identifies missing ingredients with smart substitutions

Perfect for home cooks looking to reduce food waste, try new cuisines, or find inspiration when staring at a fridge full of random ingredients.

Run Your .habit File

Run on Mobile

[ ] Download the Cortex App from store or the downloads page
[ ] Open the Cortex App on your device
[ ] Tap "Open Habit" or "+" button
[ ] Select your .habit file from your device storage
[ ] The habit will be loaded and ready to run

Run on Desktop

[ ] Download the Cortex App for your platform from the downloads page
[ ] Install and open the Cortex App
[ ] Click "Open Habit" or drag & drop your .habit file
[ ] The habit will be loaded and ready to run
[ ] Optional: Place a .env file in the same directory as your .habit file to override environment variables

Run on Server

Run your .habit file as a server using the Cortex CLI:

bash

# Install and run in one command
npx @ha-bits/cortex --config ./your-app.habit

[ ] Make sure Node.js 20+ is installed
[ ] Run the command above with your .habit file path
[ ] Server will start on the specified port (default: 3000)
[ ] Access the app at http://localhost:3000
[ ] Optional: Place a .env file next to your .habit file - it will automatically override any embedded environment variables

Run Serverless

For serverless or containerized deployments, we recommend using Docker:

bash

# Using Docker (recommended for serverless)
docker run -p 3000:3000 -v $(pwd)/your-app.habit:/app/habit.habit \
  node:20-alpine npx @ha-bits/cortex --config /app/habit.habit --host 0.0.0.0

Or create a Dockerfile:

dockerfile

FROM node:20-alpine
WORKDIR /app
COPY your-app.habit ./
COPY .env ./ # Optional: include environment variables
RUN npm install -g @ha-bits/cortex
EXPOSE 3000
CMD ["cortex", "--config", "./your-app.habit", "--host", "0.0.0.0"]

[ ] Create a Dockerfile or use the Docker run command above
[ ] Deploy to your preferred cloud provider (AWS, GCP, Azure, etc.)
[ ] Configure environment variables via your cloud provider's secrets management
[ ] Set up health checks at /habits/base/api endpoint

Workflow Visualization

Requirements

OpenAI API key (for GPT-4o and DALL-E 3)
SQLite (auto-configured)

Key Files

stack.yamlgenerate-recipe.yamlget-recipe.yamllist-recipes.yaml

yaml

version: "1.0"

workflows:
  - id: generate-recipe
    path: ./habits/generate-recipe.yaml
    enabled: true

  - id: list-recipes
    path: ./habits/list-recipes.yaml
    enabled: true

  - id: get-recipe
    path: ./habits/get-recipe.yaml
    enabled: true

server:
  port: 13000
  host: "0.0.0.0"
  frontend: ./frontend
  openapi: true

logging:
  level: info
  outputs: [console]
  format: text
  colorize: true

yaml

id: generate-recipe
name: Generate Recipe
description: Generate a recipe based on available ingredients and dietary preferences

input:
  - id: ingredients
    type: string
    required: true
    description: Comma-separated list of available ingredients
  - id: restrictions
    type: string
    required: false
    description: Dietary restrictions (vegetarian, vegan, gluten-free, etc.)
  - id: cuisine
    type: string
    required: false
    description: Preferred cuisine type
  - id: mealType
    type: string
    required: false
    description: breakfast, lunch, dinner, snack, dessert
  - id: servings
    type: number
    required: false
    description: Number of servings

nodes:
  # Generate the recipe
  - id: generate-recipe
    type: bits
    data:
      framework: bits
      source: npm
      module: "@ha-bits/bit-intersect"
      resource: ask_chatgpt
      operation: ask_chatgpt
      credentials:
        intersect:
          host: "{{habits.env.HABITS_INTERSECT_HOST}}"
          apiKey: "{{habits.env.HABITS_INTERSECT_API_KEY}}"
      params:
        model: gpt-4o
        cast: true
        temperature: 0.8
        maxTokens: 2000
        systemMessage: |
          You are a professional chef and recipe developer. Create delicious, practical recipes
          that make the most of available ingredients. Always return valid JSON.
        prompt: |
          Create a recipe using these ingredients: {{habits.input.ingredients}}
          
          Dietary restrictions: {{habits.input.restrictions}}
          Cuisine preference: {{habits.input.cuisine}}
          Meal type: {{habits.input.mealType}}
          Servings: {{habits.input.servings}}
          
          Return a JSON object with this structure:
          {
            "title": "Recipe name",
            "description": "Brief appetizing description",
            "prepTime": "15 minutes",
            "cookTime": "30 minutes",
            "totalTime": "45 minutes",
            "servings": 4,
            "difficulty": "easy/medium/hard",
            "cuisine": "Italian/Mexican/etc",
            "tags": ["quick", "healthy", "comfort food"],
            "ingredients": [
              {"item": "ingredient name", "amount": "1 cup", "notes": "optional prep notes"}
            ],
            "instructions": [
              {"step": 1, "instruction": "Step description", "tip": "optional pro tip"}
            ],
            "tips": ["Chef's tips for best results"],
            "variations": ["Alternative versions or substitutions"],
            "pairings": ["What to serve with this dish"]
          }
          
          Make the recipe creative but practical. Include helpful tips.
          Return ONLY the JSON object.

  # Generate recipe image
  - id: generate-image
    type: bits
    data:
      framework: bits
      source: npm
      module: "@ha-bits/bit-intersect"
      resource: generate_image
      operation: generate_image
      credentials:
        intersect:
          host: "{{habits.env.HABITS_INTERSECT_HOST}}"
          apiKey: "{{habits.env.HABITS_INTERSECT_API_KEY}}"
      params:
        model: gpt-image-1
        prompt: "Professional food photography of {{generate-recipe.title}}, plated beautifully, natural lighting, appetizing, high-end restaurant presentation"
        resolution: 1024x1024
        quality: medium

  # Generate shopping list for missing ingredients
  - id: shopping-list
    type: bits
    data:
      framework: bits
      source: npm
      module: "@ha-bits/bit-intersect"
      resource: ask_chatgpt
      operation: ask_chatgpt
      credentials:
        intersect:
          host: "{{habits.env.HABITS_INTERSECT_HOST}}"
          apiKey: "{{habits.env.HABITS_INTERSECT_API_KEY}}"
      params:
        model: gpt-4o
        temperature: 0.3
        maxTokens: 500
        prompt: |
          Compare these available ingredients: {{habits.input.ingredients}}
          
          With these required ingredients from the recipe:
          {{habits.input.ingredients}}
          
          List any ingredients the user needs to buy. Return as JSON:
          {
            "needed": [{"item": "ingredient", "amount": "quantity"}],
            "optional": [{"item": "ingredient", "reason": "for garnish/extra flavor"}],
            "substitutions": [{"missing": "item", "substitute": "alternative", "notes": "how to substitute"}]
          }
          
          Return ONLY the JSON object.

  # Calculate estimated nutrition
  - id: nutrition
    type: bits
    data:
      framework: bits
      source: npm
      module: "@ha-bits/bit-intersect"
      resource: ask_chatgpt
      operation: ask_chatgpt
      credentials:
        intersect:
          host: "{{habits.env.HABITS_INTERSECT_HOST}}"
          apiKey: "{{habits.env.HABITS_INTERSECT_API_KEY}}"
      params:
        model: gpt-4o
        temperature: 0.2
        maxTokens: 400
        prompt: |
          Estimate the nutritional information per serving for this recipe:
          {{generate-recipe}}
          
          Return as JSON:
          {
            "perServing": {
              "calories": 350,
              "protein": "25g",
              "carbs": "30g",
              "fat": "15g",
              "fiber": "5g",
              "sodium": "500mg"
            },
            "healthNotes": ["High in protein", "Good source of fiber"],
            "dietaryInfo": ["gluten-free", "dairy-free", etc if applicable]
          }
          
          Be reasonable with estimates. Return ONLY the JSON.

  # Save recipe
  - id: save-recipe
    type: bits
    data:
      framework: bits
      source: npm
      module: "@ha-bits/bit-database"
      resource: insert
      operation: insert
      params:
        collection: "recipes"
        document:
          recipe: "{{generate-recipe}}"
          image: "{{generate-image}}"
          shoppingList: "{{shopping-list}}"
          nutrition: "{{nutrition}}"
          inputIngredients: "{{habits.input.ingredients}}"
          restrictions: "{{habits.input.restrictions}}"
          createdAt: "{{habits.now}}"

edges:
  - source: generate-recipe
    target: generate-image
  - source: generate-recipe
    target: shopping-list
  - source: generate-recipe
    target: nutrition
  - source: generate-image
    target: save-recipe
  - source: shopping-list
    target: save-recipe
  - source: nutrition
    target: save-recipe

output:
  recipeId: "{{save-recipe.id}}"
  recipe: "{{generate-recipe}}"
  image: "{{generate-image}}"
  shoppingList: "{{shopping-list}}"
  nutrition: "{{nutrition}}"

yaml

id: get-recipe
name: Get Recipe
description: Get a specific recipe by ID

input:
  - id: id
    type: string
    required: true

nodes:
  - id: fetch-recipe
    type: bits
    data:
      framework: bits
      source: npm
      module: "@ha-bits/bit-database"
      resource: findOne
      operation: findOne
      params:
        collection: "recipes"
        filter: '{"_id": "{{habits.input.id}}"}'

edges: []

output:
  recipe: "{{fetch-recipe.document}}"
  found: "{{fetch-recipe.found}}"

yaml

id: list-recipes
name: List Recipes
description: Get all saved recipes

input:
  - id: limit
    type: number
    required: false

nodes:
  - id: query-recipes
    type: bits
    data:
      framework: bits
      source: npm
      module: "@ha-bits/bit-database"
      resource: query
      operation: query
      params:
        collection: "recipes"
        limit: "{{habits.input.limit}}"
        sort: '{"createdAt": -1}'

edges: []

output:
  recipes: "{{query-recipes.results}}"
  count: "{{query-recipes.count}}"

Quick Start

Run using the Habits CLI wrapper, recommended if you develop local Habits

# First, download the example files
npx habits@latest cortex --config ./ai-cookbook/stack.yaml

Download ai-cookbook.habit

AI Cookbook ​

Features at a Glance ​

Run Your .habit File ​

Run on Mobile

Run on Desktop

Run on Server

Run Serverless

Workflow Visualization ​

Requirements ​

Key Files ​

Quick Start ​

AI Cookbook

Features at a Glance

Run Your .habit File

Workflow Visualization

Requirements

Key Files

Quick Start