LLM Recipe Parsing for Ingredient-to-Item Database Linking

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Online recipes often use generic terminology for ingredients, making it difficult to link them to specific products in an online database, which complicates the process of obtaining ingredients for users.

Innovation Solution

Employing natural language processing and machine learning technologies to map unstructured text content in recipes to structured data, enabling the identification of specific ingredients and their corresponding items in an online database, thereby facilitating the creation of linked recipes that can be easily searched and ordered.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If manual methods are used to add recipes to online platforms, then recipe content can be added, but the process becomes extremely labor intensive

Engineering Contradiction:

Improverecipe addition efficiencyVSAvoidtime for manual recipe entry

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The patent replaces manual mechanical data entry with automated machine learning models and natural language processing systems. These AI systems automatically extract ingredient information from unstructured recipe text and images, converting it into structured database entries without human intervention, thereby dramatically improving productivity and eliminating time loss associated with manual entry

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables recipes to self-process through automated ML pipelines that independently extract, structure, and store recipe data. The machine learning models autonomously handle the entire recipe ingestion workflow from raw unstructured data to structured database records, making the system self-sufficient and eliminating the need for manual labor

Inventive Principle:

Principle #25Self-service

2Adaptability or versatility

If generic terminology is used for ingredients in recipes, then recipes can be created with simple language, but the ingredients cannot be linked to specific products in the database

Engineering Contradiction:

Improverecipe language flexibilityVSAvoidingredient identification precision

Core Design Contradiction:

Adaptability or versatilityVSLoss of information

Solution Approach 1:

The patent introduces machine learning models and natural language processing systems as intermediaries between generic ingredient descriptions and specific database products. These intermediary systems analyze the context, quantity, and characteristics of ingredients mentioned in recipes and map them to corresponding structured database entries, preserving both the flexibility of natural language and the precision of product identification

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system transforms ingredient descriptions by adding structural parameters (quantity, unit, ingredient type) to generic text. The machine learning models enrich unstructured ingredient mentions with additional attributes and metadata, converting simple text like 'milk' into structured data with specific product identifiers, quantities, and nutritional information, thereby maintaining language simplicity while gaining identification precision

Inventive Principle:

Principle #35Parameter changes

3Ease of operation

If unstructured text content is used to describe ingredients, then recipes can be written naturally, but it becomes impossible to represent recipes in terms of items available from an online database

Engineering Contradiction:

Improverecipe creation simplicityVSAvoiddata structure complexity

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The patent replaces complex manual data structuring operations with automated machine learning systems. These systems automatically parse unstructured recipe text, extract ingredient information, and generate structured database records with proper formatting, relationships, and metadata, thereby maintaining ease of recipe creation while managing data structure complexity through automation

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system extracts structured ingredient data from unstructured recipe text using natural language processing and machine learning. The extraction process isolates relevant information (ingredient names, quantities, units) from the natural language description and transforms it into structured database entries, preserving the simplicity of natural language input while creating organized, queryable data structures

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS20260065347A1Parsing Text Content to Generate Links to Database of Items Using Large Language Models

Publication Date: 2026.03.05 MAPLEBEAR INC

AI summary

Item linked recipe generation using machine learning is described. Raw data is received that describes a recipe that uses ingredients. Ingredient descriptors are extracted from the raw data for the ingredients. Parsed ingredient data is determined using the ingredient descriptors and a large language model, such that the parsed ingredient data for each ingredient includes a name, a quantity, and a unit of measure. The name of each ingredient is mapped to a corresponding ingredient identifier that is part of an ingredient database. And each ingredient identifier in the ingredient database is associated with a corresponding item that is available for sale at one or more sources. A linked recipe is generated that includes for each ingredient: an ingredient identifier, a quantity of the ingredient, and a unit of measure of the quantity. A recommendation for the linked recipe is provided to a user client device.