I wanted to share with the group the way I am making custom recipes in the Levels app. I’ll use the example of a custom salad I frequently eat for lunch, and list the ingredients shortly. I use ChatGPT and Excel to streamline this process to get the “bottom line” nutritional list to put into the Levels app.

If you don’t have a ChatGPT subscription, free access is available at chat.openai.com.

Once I have the list of ingredients for any meal or custom food I want to make, I instruct ChatGPT like this:

"If I make a salad with these ingredients: 2/3 cup romaine Lettuce, 1 teaspoon sunflower seeds, one Kalamata olive, 1 tablespoon avocado, 1 tablespoon fermented mustard, 1 tablespoon chopped artichoke, 1/2 cup cottage cheese, 1 teaspoon shredded Parmesan cheese, and 1.5 tablespoons avocado mayonnaise, please list the nutritional values in a table for these items:

Calories, Total Fat (g), Saturated Fat (g), Trans Fat (g), Cholesterol (mg), Sodium (mg), Total Carbs (g), Fiber (g), Total Sugars (g), Protein (g), Vitamin D (IU), Calcium (mg), Iron (mg), and Potassium (mg)."

The first paragraph is my instruction and list of ingredients for the salad. The second paragraph is the list of values I want (which is the standard list of nutritional items on any food package, and in the same order you enter them in the Levels app).

The table ChatGPT generates can be downloaded as a CSV file and opened in Excel. (I just click the down arrow in ChatGPT to download the table.) You can find my sample salad file at this link:

Once you open the CSV file, you can see the list of ingredients down the left column and the nutritional values across the top. This was all provided by ChatGPT based on my instructions. To keep it simple, I didn’t ask ChatGPT to total the values (I could have asked). Instead, I calculate the totals inside the CSV file by adding formulas. Then, because ChatGPT sometimes makes mistakes, I check the values to ensure they make sense. (In doing this, you might also check the ingredient list on the back of the food you’re using. For example, ChatGPT gave me values for cottage cheese that surprised me, so I checked the container label. ChatGPT was correct.) As you can see in the CSV file, I added a =sum() formula on each column. This is on row 11.

Some users suggest going to the Levels app, selecting "Food," then "Describe," and entering your recipe ingredients and quantities to let the Levels AI calculate the nutritional info. I tried that several times, and unfortunately, it’s almost always incorrect. The method I described above is more work but bullet-proof. Once you get the hang of it, it can be done quickly.

By the way, if you make a custom recipe for low-carb muffins, for example, and give ChatGPT the list of recipe ingredients, after you get the CSV file and total the values, you need to divide them by the number of muffins the recipe yields. For example, if the recipe yields 12 muffins, divide the total values by 12 to get the nutritional information per muffin.

Hope this helps.