What is the CrawlFeeds recipe dataset?

The CrawlFeeds recipe dataset contains over 300,000 structured recipes sourced from AllRecipes, BBC Food, and Taste of Home. Each recipe includes 28 parsed data fields: ingredients with metric conversions, full nutrition data (calories, macros, vitamins, % daily values), cook and prep times, cuisine type, diet tags, step-by-step instructions, and high-resolution image URLs. Available in CSV and JSON formats for download.

How does CrawlFeeds recipe dataset compare to RecipeNLG or Kaggle?

Unlike free research datasets such as RecipeNLG (1M recipes, non-commercial only), Food.com on Kaggle (raw text), or the MIT Recipe1M+ (research-licensed), the CrawlFeeds recipe dataset is commercially licensed, actively maintained, and delivers structured parsed data including metric-converted ingredients and granular nutrition breakdowns per serving. It is ready for production use in apps, AI models, and commercial analytics without legal restrictions on commercial deployment.

Can I use this recipe dataset for AI and machine learning training?

Yes. The CrawlFeeds recipe dataset is widely used for training NLP models for recipe generation and ingredient substitution, fine-tuning recommendation engines, building nutrition-aware meal planners, and training food image recognition models. The structured JSON format with parsed ingredients, cuisine labels, and diet tags makes it directly compatible with standard ML pipelines in Python and R.

How many recipes are in this dataset?

The dataset contains over 300,000 recipes from trusted sources including AllRecipes.com, BBC Food, and Taste of Home, with 28 structured data points per recipe.

What data points are included for each recipe?

Each recipe includes 28 data fields: name, source URL, raw ingredients, parsed ingredients (quantity, unit, metric conversion), cuisine, country, region, diet types (vegetarian, vegan, keto), description, prep time, cook time, total time in ISO 8601 format, publish date, high-res image URLs, category, yield, author, keywords, step-by-step instructions, and full parsed nutrition data including calories, macros, vitamins, minerals, and % daily values.

What nutrition information is available per recipe?

Comprehensive per-serving nutrition data: calories, total fat, saturated fat, cholesterol, sodium, total carbohydrates, dietary fiber, sugars, protein, vitamins (C, A, etc.), minerals (calcium, iron, potassium), and % daily values. Available in both structured JSON and human-readable text.

What export formats are available for the recipe dataset?

The recipe dataset is available in CSV and JSON formats. CSV works with Excel, Google Sheets, pandas, and SQL imports. JSON preserves nested structures for parsed ingredients and nutrition data, ideal for direct API and app integration.

Can I filter recipes by cuisine or diet type?

Yes. Recipes are tagged with cuisine (Japanese, Brazilian, Italian, American, etc.), country, region, and diet types (vegetarian, non-vegetarian, vegan, keto, gluten-free). These fields allow direct filtering and segmentation in your application or ML pipeline.

Is the recipe dataset available for commercial use?

Yes. The CrawlFeeds recipe dataset is available for commercial use under the CrawlFeeds terms of service. Each record includes source attribution information. Unlike free research alternatives (RecipeNLG, Recipe1M+) which restrict commercial use, CrawlFeeds data is licensed for production and commercial applications.

Recipe Datasets - 300K+ Recipes with Nutrition &

300K+ Structured Recipe Data Collection

Our comprehensive recipe dataset contains over 300,000 recipes from trusted sources like AllRecipes.com, with 28 structured data points per recipe. Each recipe includes parsed ingredients with automatic metric conversions, comprehensive nutrition data, cooking times, cuisine classifications, and step-by-step instructions.

Perfect for AI training, meal planning apps, nutrition analysis, recipe recommendation engines, and culinary innovation. Data is available in CSV and JSON formats, ready to integrate into your applications.

Download Dataset Browse All Datasets

28 Structured Data Points Per Recipe

Every recipe is meticulously parsed and structured with comprehensive metadata

Basic Information

Name, recipe URL, description, author, source, and published date for full attribution and context.

Parsed Ingredients

Structured JSON with quantity, unit, ingredient name, and automatic metric conversions (cups to ml, oz to g).

Nutrition Data

Calories, macros, vitamins, minerals with % daily values in both structured and readable formats.

Timing Information

Prep time, cook time, and total time in ISO 8601 format (PT40M) for easy parsing.

Classification & Tags

Cuisine type (Japanese, Brazilian), country, region, diet types (vegetarian, vegan, keto), and category.

Media & Instructions

High-res images, gallery arrays, step-by-step instructions, recipe yield, and keywords.

Sample Recipe Records

Below are 5 sample records from our 300K+ recipe dataset showing the depth and quality of structured data

Name	Cuisine / Category	Times	Yield	Ingredients Count	Nutrition (Per Serving)	Source
Homemade Pickled Ginger (Gari) Pickled ginger, or gari, is served as a palate cleanser alongside sushi or sashimi.	Japanese Condiment	Prep: 40m Cook: 5m Total: 45m	32 servings	4 ingredients All parsed with metric conversion	14 kcal 3g Carbs 83mg Sodium	AllRecipes
Pork and Black Bean Stew Brazilian pork and black bean stew with chorizo sausage and spices.	Brazilian Dinner Non-Vegetarian	Prep: 20m Cook: 40m Total: 60m	4 servings	10 ingredients All parsed with metric conversion	482 kcal 33g Protein 27g Fat	AllRecipes
Cookie Press Shortbread Traditional spritz cookies made with butter, sugar, and cornstarch.	Dessert	Prep: 25m Cook: 10m Total: 35m	24 cookies	5 ingredients All parsed with metric conversion	116 kcal 11g Carbs 8g Fat	AllRecipes
Jalapeno Poppers of Champions Jalapeno halves filled with cream cheese, wrapped in bacon.	U.S. Appetizer Vegetarian Option	Prep: 15m Cook: 12m Total: 30m	18 servings	4 ingredients All parsed	73 kcal 7g Fat 2g Protein	AllRecipes
Lobster Colorado Bacon-wrapped filet mignons topped with lobster meat.	New England Dinner Non-Vegetarian	Prep: 10m Cook: 30m Total: 40m	4 steaks	7 ingredients All parsed with metric conversion	585 kcal 52g Protein 41g Fat	AllRecipes

Note: Each recipe includes the full ingredient list with quantities and units, step-by-step instructions, complete nutrition breakdown with vitamins and minerals, high-resolution images, and more. The table above shows a simplified view of the 28 data points available per recipe.

Explore Recipe Datasets by Source

🍲 AllRecipes Recipe Dataset

Comprehensive collection with diverse cuisines and detailed nutrition.

View Dataset

🏡 Taste of Home Recipes

Family-friendly recipes with home cooking focus and ratings.

View Dataset

📺 BBC Food Recipes

Professional recipes from BBC chefs with international flavors.

View Dataset

Use Cases for Recipe Datasets

→
AI & Machine Learning
Train NLP models for recipe generation, ingredient substitution, and cooking assistants.
→
Meal Planning Apps
Power personalized meal plans with filterable recipes by diet, cuisine, and cooking time.
→
Nutrition Analysis
Access detailed nutrition data with calories, macros, vitamins, and % daily values.
→
Recipe Recommendation Engines
Build smart recommendations using cuisine types, diet tags, and ingredient similarities.
→
Food Analytics & Research
Analyze cooking trends, popular ingredients, and nutritional patterns across cuisines.

300K+

Total Recipes

28

Data Points

100%

Parsed Ingredients

Multiple

Cuisines

Supported Data Formats

Data is available in multiple formats for easy integration with your applications.

CSV JSON Structured Data

Frequently Asked Questions

Our recipe dataset contains over 300,000 recipes from trusted sources like AllRecipes.com, with structured data across 28 distinct fields per recipe.

Each recipe includes 28 data points: name, URL, raw ingredients, parsed ingredients (with quantity, unit, metric conversion), cuisine, country, region, diet types, description, cook/prep/total time, publish date, images, category, yield, author, keywords, step-by-step instructions, parsed nutrition data (calories, macros, vitamins, minerals with % daily values), source, scrape timestamp, and version info.

Yes! Every recipe has both raw ingredients and parsed ingredients in structured JSON format. Parsed data includes quantity (numeric), unit (cup, tsp, oz, etc.), ingredient name, and automatic metric conversions (ml, g). For example: '1 cup butter' is parsed as 1.0 cup = 240ml.

Comprehensive nutrition data per serving: calories, total fat, saturated fat, cholesterol, sodium, total carbohydrates, dietary fiber, sugars, protein, vitamins (C, etc.), minerals (calcium, iron, potassium), and % daily values. Available in both structured JSON and human-readable text formats.

Absolutely! Recipes are tagged with cuisine (Japanese, Brazilian, Italian, etc.), country, region, and diet types (vegetarian, non-vegetarian, vegan, keto, etc.). You can easily filter and segment the dataset based on these classifications.

Datasets are available in CSV and JSON formats. CSV is perfect for Excel/spreadsheet analysis, while JSON preserves the complex nested structures of parsed ingredients and nutrition data.

Yes! Each recipe includes high-resolution image URLs (typically 1500px width from AllRecipes). Some recipes also include gallery arrays for multiple images.

Times are in ISO 8601 duration format: PT40M (40 minutes prep), PT5M (5 min cook), PT45M (45 min total). This machine-readable format is easy to parse and convert to any display format.

Yes, the dataset is available for commercial use. Each record includes proper source attribution (source domain, author, URL) for transparency and compliance.

All recipes include scraped_at timestamp (when collected) and date_published (original publish date). Most recent data collection was March 2026. The dataset is updated regularly with new recipes and sources.

What Does "Structured" Actually Mean?

Most recipe datasets give you raw text. CrawlFeeds delivers machine-ready structured data — every field parsed, typed, and normalised so you can load it directly into your app or model without preprocessing.

Parsed Ingredients

Each ingredient string is split into four typed fields plus an automatic metric conversion, so your app never needs to parse "2½ cups of all-purpose flour" itself.

Field	Type	Example
`quantity`	Float	2.5
`unit`	String	cup
`ingredient`	String	all-purpose flour
`metric_quantity`	Float	600.0
`metric_unit`	String	ml

// JSON — parsed ingredients example

{
  "raw": "2½ cups all-purpose flour",
  "quantity": 2.5,
  "unit": "cup",
  "ingredient": "all-purpose flour",
  "metric_quantity": 600.0,
  "metric_unit": "ml"
}

Full Nutrition Per Serving

Every recipe carries a complete per-serving nutrition breakdown — not just calories. Vitamins, minerals, and % daily values are included, parsed from the source and structured in both JSON and human-readable formats.

Macros: Calories, total fat, saturated fat, cholesterol, sodium, total carbs, dietary fibre, sugars, protein

Vitamins: Vitamin C, Vitamin A, Vitamin D, and others where present

Minerals: Calcium, iron, potassium, magnesium

% Daily Values: All nutrients include % DV relative to a 2,000 kcal diet

// JSON — nutrition per serving

{
  "calories": 482,
  "total_fat_g": 27,
  "saturated_fat_g": 9,
  "cholesterol_mg": 112,
  "sodium_mg": 680,
  "total_carbs_g": 18,
  "dietary_fiber_g": 4,
  "sugars_g": 2,
  "protein_g": 33,
  "vitamin_c_pct": 12,
  "calcium_pct": 6,
  "iron_pct": 18,
  "potassium_mg": 742
}

Classification & Timing

All time fields are stored in ISO 8601 duration format so any date library can parse them instantly. Cuisine, diet, and category tags are normalised strings — no cleaning required.

Field	Format / Example
`prep_time`	ISO 8601 — `PT25M`
`cook_time`	ISO 8601 — `PT40M`
`total_time`	ISO 8601 — `PT65M`
`cuisine`	String — `Japanese`, `Brazilian`
`diet_types`	Array — `["vegetarian","keto"]`
`category`	String — `Dessert`, `Dinner`
`yield`	String — `4 servings`, `24 cookies`

Images & Attribution

High-resolution image URLs (typically 1500 px wide) are included per recipe. Every record also carries full source attribution so you can display provenance in your product.

// JSON — image & source fields

{
  "image": "https://cdn.allrecipes.com/...",
  "gallery": ["https://...img1", "https://...img2"],
  "source": "allrecipes.com",
  "author": "John Doe",
  "source_url": "https://allrecipes.com/recipe/...",
  "date_published": "2023-08-14",
  "scraped_at": "2026-03-01T08:12:44Z"
}

All 28 Fields at a Glance

Basic Fields

`name`	String
`description`	Text
`author`	String
`source`	String
`source_url`	String
`date_published`	Date
`scraped_at`	Datetime
`version`	String

Ingredients Fields

`ingredients_raw`	Array
`ingredients_parsed`	Array of objects
`ingredient_count`	Integer

Nutrition Fields

`calories`	Integer
`total_fat_g`	Float
`saturated_fat_g`	Float
`cholesterol_mg`	Integer
`sodium_mg`	Integer
`total_carbs_g`	Float
`dietary_fiber_g`	Float
`sugars_g`	Float
`protein_g`	Float
`vitamins`	Object
`minerals`	Object

Metadata Fields

`prep_time`	ISO 8601
`cook_time`	ISO 8601
`total_time`	ISO 8601
`recipe_yield`	String
`cuisine`	String
`diet_types`	Array
`category`	String
`keywords`	Array
`image`	String
`gallery`	Array
`instructions`	Array

Ready to Get Started?

Download our 300K+ recipe dataset or browse other food & beverage datasets.

Download Dataset Browse More

Powerful Recipe Datasets

300K+ Structured Recipe Data Collection

28 Structured Data Points Per Recipe

Basic Information

Parsed Ingredients

Nutrition Data

Timing Information

Classification & Tags

Media & Instructions

Sample Recipe Records

Explore Recipe Datasets by Source

🍲 AllRecipes Recipe Dataset

🏡 Taste of Home Recipes

📺 BBC Food Recipes

Use Cases for Recipe Datasets

300K+

28

100%

Multiple

Supported Data Formats

Frequently Asked Questions

Further Reading on Recipe Intelligence

What Does "Structured" Actually Mean?

Parsed Ingredients

Full Nutrition Per Serving

Classification & Timing

Images & Attribution

All 28 Fields at a Glance

Ready to Get Started?

Powerful Recipe Datasets

300K+ Structured Recipe Data Collection

28 Structured Data Points Per Recipe

Basic Information

Parsed Ingredients

Nutrition Data

Timing Information

Classification & Tags

Media & Instructions

Sample Recipe Records

Explore Recipe Datasets by Source

🍲 AllRecipes Recipe Dataset

🏡 Taste of Home Recipes

📺 BBC Food Recipes

Use Cases for Recipe Datasets

300K+

28

100%

Multiple

Supported Data Formats

Frequently Asked Questions

How many recipes are in this dataset?

What data points are included for each recipe?

Are ingredients automatically parsed with quantities?

What nutrition information is available?

Can I filter recipes by cuisine or diet type?

What export formats are available?

Are recipe images included?

How is cooking time formatted?

Can I use this data commercially?

How fresh is the recipe data?

Further Reading on Recipe Intelligence

What Does "Structured" Actually Mean?

Parsed Ingredients

Full Nutrition Per Serving

Classification & Timing

Images & Attribution

All 28 Fields at a Glance

Ready to Get Started?