Recipe Datasets for AI, Apps, and Food Research

Crawl Feeds provides source-specific recipe datasets for teams that need structured recipe records for applications, analytics, or research. Its recipe listings include collections such as Food.com, BBC Food, and Allrecipes, with fields that can include recipe titles, ingredients, instructions, timings, categories, nutrition, and source URLs. The fields and record counts vary by dataset, so compare each listing and inspect its sample before selecting a collection.

Recipe data for product teams and researchers

Recipe data is useful when a product or analysis needs more than individual recipe pages. Structured records can support recipe discovery, ingredient-based search, recommendation prototypes, food trend analysis, and culinary research. Crawl Feeds’ catalog offers recipe datasets collected from named source websites; each source collection has its own coverage and schema. Choose by the fields, sources, and scope your project actually needs—not by a single headline record count.

Browse the Crawl Feeds recipe dataset catalog

Explore Crawl Feeds’ web data collection service

What recipe datasets are available from Crawl Feeds?

Crawl Feeds’ indexed catalog and dataset pages currently describe several source-specific recipe collections. Examples include a Food.com collection listed at nearly 300,000 recipes, a BBC Food collection listed at about 9,700 recipes, and an Allrecipes collection listed at 48,520 recipes. These are listing figures and can change; check the linked dataset page for the current count and scope.

Which Crawl Feeds recipe dataset should you choose?

Choose the collection whose documented source coverage and fields fit the job. For broad recipe analysis, compare the size and source scope of Food.com, BBC Food, and Allrecipes collections. For ingredient search or recipe parsing, verify that ingredient values are supplied in the representation your pipeline needs. For nutrition-related analysis, inspect the nutrition fields and missingness in a sample rather than assuming every recipe has complete nutrition data. For reviews or audience sentiment, select a reviews dataset only if review records—not just recipes—are required.

How to evaluate a recipe dataset before purchase

• Open the specific dataset listing and compare its source, current record count, schema, and delivery format with your requirements.
• Use the available sample to inspect actual records, data types, missing values, ingredient structure, and instruction quality.
• Check whether you need raw ingredients, parsed ingredients, nutrition, images, ratings, author/source URLs, or timestamps; availability differs by collection.
• Ask how often the collection is refreshed if freshness matters. Do not infer a refresh schedule from a scrape timestamp in an example record.
• Review the applicable dataset terms and source-specific rights for your intended use, including commercial use, model development, redistribution, and attribution.
• If the catalog does not match your source, fields, filters, or refresh needs, ask Crawl Feeds whether a custom collection or recurring feed can be scoped.

How teams use structured recipe data

• Recipe apps: index titles, ingredients, categories, and instructions to support discovery and filtering.
• Recommendation prototypes: combine recipe attributes with your own interaction data where available; recipe records alone do not necessarily include user-preference history.
• Food trend analysis: examine ingredients, cuisines, categories, and publication or collection dates while accounting for source coverage and collection bias.
• Research and machine-learning experiments: test tasks such as recipe classification or generation only after reviewing provenance, fields, and dataset terms.
• Nutrition workflows: treat listed nutrition as source data that needs validation for the intended use; recipe records are not individualized health advice.

Ready-made dataset or custom collection?

A ready-made Crawl Feeds dataset is a practical starting point when an available source and schema meet your needs. If you need a different site, selected fields, a specific filter, or recurring updates, Crawl Feeds also describes custom extraction and recurring-feed options across its data services. Confirm feasibility, delivery schedule, format, price, and terms with Crawl Feeds for your particular request; they are not guaranteed for every recipe source or project.

Request a custom data solution

How are research recipe datasets different?

Research datasets such as RecipeNLG and Recipe1M+ are published in academic contexts for particular research tasks. RecipeNLG focuses on semi-structured recipe text generation; Recipe1M+ concerns cross-modal learning with recipe and food-image data. They are not interchangeable with a source-specific commercial data collection. Compare their original papers and access terms with the fields, source coverage, format, and service requirements of the Crawl Feeds listing you are considering.