Genesant.AI provides enterprise search, recognition, and recommendation SaaS services for personal health management using natural language understanding (NLU) AI models. Their technology leverages proprietary data from billions of customer interactions to power services for media companies, manufacturers, and institutional service providers.
Funding
Funding not disclosed
Founders
Product
Problem
Tracking diets, ordering meals, and searching for nutritional information often requires users to manually input and interpret unstructured, conversational descriptions of food. This process is time-consuming and prone to errors, hindering efficient personal health management.
Solution
Genesant AI offers a suite of natural language understanding (NLU) AI models that extract structured data from conversational descriptions of food, enabling developers to build applications for diet tracking, meal ordering, and nutritional information retrieval. The platform's API identifies foods, displays nutritional content, and retrieves UPC identifiers from user inputs, streamlining the process of logging and tracking individual diets. By leveraging a pipeline of neural networks, Genesant AI transforms unstructured text into actionable data, facilitating quicker and easier access to relevant information. The platform's capabilities extend to splitting food lists, tagging food components, quantifying ingredients, and matching descriptions to a comprehensive food database.
Target Audience
The primary target audience includes developers in the health and nutrition industries building applications for diet tracking, meal planning, grocery shopping, and nutritional information retrieval.
Features
- Recognition API accessible via REST and GraphQL interfaces
- Suite of AI models including Splitter, Tagger, Quantizer, Matcher, and Recognizer
- Splitter: Semantic parser using temporal convolutional networks (TCN) and BERT-based transformer networks to identify food boundaries in lists
- Tagger: Semantic role annotator using a BERT-based transformer network to label individual words within food descriptions
- Quantizer: Quantity and unit extractor/classifier using a transformer network to identify scalar magnitudes and unit classes
- Matcher: Extreme multiclass classifier using TCN and BERT-based transformer networks to identify foods from conversational input
- Access to a knowledge base of over 250,000 foods