Use our state of the art image recognition models to analyze what your users eat, determine what's trending and make your content easier to find.
Detect high level categories like beverage and soup as well as specific dishes and ingredients.
Predict item attributes like cuisine, preparation method and flavor.
Get unmatched recognition accuracy with models that use a user's eating habits to refine predictions and automatically learn to recognize past meals.
Automate meal logging with a fully featured machine learning based food logging SDK.
With food recognition logging a meal takes as little as 5 seconds. Our machine learning models learn to recognize past meals and improve with usage.
Integrated with our Food Knowledge Graph that contains a large set of commonly eaten foods, with nutrition facts, and hierarchical structure.
Includes faceted instant search that's personalized to the user, so most likely foods can be found with a single key stroke.
Reason about food. Get access to our database of nutrition information, ingredients, categorizations, and relations.
Organized from high level concepts like beverage and sandwich down to specific branded products and variants.
Categorized by dietary trends, cuisine, preparation method and restrictions.
Most items include calories, macro and micro nutrients as well as ingredients.
Statistics on popularity, most likely consumption time and most likely items to be consumed with.
Collect nutrition information from any package using tailored optical character recognition models.
Unlike a standard OCR solution, our models are optimized for nutrition labels, leading to much higher accuracy.
Returns structured nutrition fact responses that are compatible with out knowledge graph representation.
Turn unstructured text into a semantic representation that's easier to index and search.
Parse food names and recipes, identifying items, preparation methods, brands and serving sizes.
Predict cuisine, dish types, high-level categories and ingredients from text.
Make sense of user generated text by find semantically similar items in your database.
Make your content easier to find by enabling users to search by image.
Help users find exactly what they're looking for instead of hoping that they'll be able to put together a specific enough text based query.
Surface more results by match similar items when exact results are not available.
View what your users are actually looking for to get deeper insights on their intent.
We're committed to your success and will work with you to provide an optimal solution for your problem.
If you need to detect or parse specific concepts we can help you tune our models for your use case.
We can help you deploy your models for bulk, offline or low latency inference, including on premise and on mobile devices.
We can help you get from 0 to 1 and make the most of our technology with hands on integration support.