We now have glycemic index and glycemic load data for many ingredients and can compute the total glycemic load for multiple ingredients or a recipe. Check out the dedicated endpoint or just find the data that is now added to the "nutrition" object called "properties", e.g.
"nutrition": { "properties": [ { "title": "Glycemic Index", "amount": 13.92, "unit": "" }, { "title": "Glycemic Load", "amount": 8.92, "unit": "" } ], }
We have updated all the widget endpoints for visualizing recipe nutrition, equipment, price breakdown, and ingredients and you can use the "play" button in the docs to preview the widgets.
Furthermore, you can find fully working example code that you can copy to your projects which we host on GitHub here: https://github.com/ddsky/spoonacular-widgets
In our last release, we introduced our new meal planning API and we were surprised how many of you adopted it so quickly! From the feedback we received we added and improved a couple endpoints:
In our last release, we introduced our new meal planning API and we were surprised how many of you adopted it so quickly! From the feedback we received we added and improved a couple endpoints:
We receive a lot of requests from users who want to build their own meal planning tools. To help you achieve this, we have added 11 new endpoints to our API, including Add/Delete from Meal Plan, Get Meal Plan Week, Generate Shopping List, and other highly requested endpoints.
These endpoints are closely connected to our meal planner on spoonacular.com. You can find out more here: https://spoonacular.com/food-api/docs#Working-with-the-Meal-Planner
We have been working on a new set of functionalities that allow you to classify food pictures! Our model is trained on over 15,000 images in 50 classes with a focus on dishes that you can get in a restaurant, such as burgers, fish and chips, and lasagna. To train our model, we picked professional photographs and pictures taken by actual restaurant visitors - see how different these two categories are?
How good is our model? After extensive training and evaluation we reached just above 90% accuracy. We also researched other classifiers and studied how well Google, Microsoft, IBM, and other companies focused on image classification work on our dataset. Here's the comparison if you're interested: https://spoonacular.com/food-api/docs#Image-Classification
We have been working on the speed of the recipe search. The new Complex Recipe Search endpoint is now five times faster.
You can now also sort recipes by a large list of criteria, such as by any macro- or micronutrient, cooking time, price, meta-score, or just random.
The nutritional information for each recipe now also contains the weight per serving:
"weightPerServing": { "amount": 269, "unit": "g" }
Additionally, the nutritional information also includes net carbs, which is defined in the API as total carbohydrates minus the fiber content.
When requesting meta-information from the autocomplete ingredient endpoint, you can now also request a list of possible units for any given ingredient.
For example, if your app allows users to enter ingredients, you might want to let them enter 3 slices of apple instead of just 1 oz of apple. The possible units tell you whether the API recognizes the unit (e.g. slices) and can compute the nutrition correctly. The API does not know "slices" for, say, peanuts, so it is beneficial to let your user only pick units that can be understood.
Information about grocery products now contains all the nutritional information we have for a product, not just macronutrients. Additionally, we added a description of the product if available and attempt to guess the supermarket aisle where you would expect to find the product.
We have published a list of the top 1,000 ingredients used in recipes and their spoonacular ingredient ids here.
All our SDKs have been recompiled, so make sure you update your SDK if you are using an older one.