Recipe Scraper API

A FastAPI service I built that scrapes recipes from any webpage (including Instagram posts), uses OpenAI's GPT-4 to extract structured data, and returns it in a standardized JSON format. The API handles various recipe formats and normalizes them into a consistent schema.

What It Does

• Scrapes recipes from any webpage, including Instagram posts
• Uses OpenAI's GPT-4 to extract structured recipe data
• Returns a well-defined JSON schema with:
  • Ingredients with quantities and units
  • Step-by-step instructions
  • Recipe metadata (prep time, cook time, difficulty, etc.)
  • Categories and descriptions

How It Works

The API uses BeautifulSoup for web scraping and OpenAI's GPT-4 for intelligent data extraction. It handles both traditional websites and Instagram posts, normalizing the data into a consistent format.

# Example of the Recipe schema
class Recipe(BaseModel):
    title: str
    description: str | None = None
    photo: str | None = None
    category: Category
    ingredients: list[Ingredient]
    instructions: list[Instruction]
    prep_time: int | None = None
    cook_time: int | None = None
    difficulty: int
    servings: int
    yield_: int | None = None
    original_link: str | None = None
    video_link: str | None = None

# Example of ingredient structure
class Ingredient(BaseModel):
    ingredient_name: str
    quantity: float
    measurement_unit: MeasurementUnit
    notes: str | None = None

Technical Bits

• Built with FastAPI for high performance
• Uses BeautifulSoup for web scraping
• Integrates with OpenAI's GPT-4 API
• Implements Pydantic models for data validation
• Handles Instagram-specific scraping
• Supports various measurement units and categories

Features

1. Universal Scraping: Works with any recipe website
2. Instagram Support: Special handling for Instagram recipe posts
3. Structured Output: Consistent JSON format for all recipes
4. Intelligent Extraction: Uses GPT-4 to understand and structure recipe data
5. Data Validation: Ensures all required fields are present and properly formatted

Why I Built It

I wanted to create a unified way to access recipe data from various sources. This API helps standardize recipe information, making it easier to work with recipe data across different platforms and formats.

Future Ideas

• Add support for more social media platforms
• Implement recipe image extraction
• Add nutritional information extraction
• Support multiple languages
• Add recipe scaling functionality
• Implement caching for frequently accessed recipes