What is Ingredient Recognition?
Ingredient recognition is a real-time recognition tool for select fresh ingredients and packaged goods you may have on hand, and will provide recipe recommendations based on those ingredients. Since it's in real-time, there is no need for additional photo taking on your end - just point the camera at your ingredients and we'll take care of the rest!
How do I use Ingredient Recognition?
No need to do anything before scanning - you should just have the ingredient clearly visible within camera view for Ingredient Recognition to be activated.
You can use it on both fresh ingredients and packaged goods (think cans, boxes, cartons, jars, etc..). Ingredient Recognition will read text on the packaging to identify products.
It works best when:
The ingredient is close enough within camera view & in a well lit condition.
The distinctive features of an ingredient are in view (For example, from the top or bottom pears and apples can look very similar, but from the side the difference in shape is clear.)
Any important text should be visible and head-on when scanning packaged goods.
If it doesn't recognize the ingredient right away but a teal box appears, try keeping the ingredient in the frame for a few seconds and it might either recognize it or prompt you to add manually.
If it's is having trouble recognizing an ingredient, try picking the ingredient up, or moving the camera to get a different view of the ingredient, especially in a way that exposes the distinctive features.
If we've partially recognized something, we will let you know what ingredient we think it is and ask you to confirm.
To add ingredients manually, use the manual add button.
If we've incorrectly recognized something, you can press and hold the ingredient tag to edit it - you'll be helping teach our model to recognize it correctly next time!