Our models are trained on millions of images and can recognize thousands of the most common foods.
Recognize commonly eaten dishes, from paninis to paellas.
Detect dish ingredients to refine results and create unlimited combinations.
Detect whole foods like apples, bananas and nuts.
We combine a lot of data, the latest deep learning models and contextual information to provide the highest accuracy.
Automatically train models to recognize previously eaten meals with human level accuracy.
Get updated suggestions with each prediction that the user confirms with our interactive input API.
Models improve for your specific use case with each image that you process.
Use our technology to train models to recognize specific foods or attributes.
Train models to recognize specific dishes with just a few example images. Improve the models even more after you deploy them.
Annotate images with custom properties and train models to predict them.
Predictions are nodes in the knowledge graph so you can reason about them.
Go up or down the relationship hierarchy to tell that "chicken" is a type of "meat" and let user's refine it to "grilled chicken" for more acurate nutritional estimates.
Most predicted items include full nutritional details, ingredients and multiple serving sizes so you can calculate caloric content.
Check categorical info such as cuisine, dish type, flavor and preparation method.
Nodes include references to external data sources such as Wikidata and USDA Nutrition Database.
👴🏻💦 Developers, Developers, Developers!
Use our JSON based REST API to process images using any programming language.
Use our fully featured clients to integrate image recognition into your app in as little as 4 lines of code.
Use our customizable components to speed up UI integration for common user flows.
response = biteai.recognize(image)
> "Hot Dog"
Get a better picture of your users by enabling them to quickly track what they're eating.
Analyze your images to expand your search index with higher level concepts. Keep users engaged with similar item recommendations.
Verify that user submitted images match the description and group existing images by dish or ingredients.
Make sense of unstructured data at scale by Analyze data at scale and understand trends at a deeper level.
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.