Personal Soylent Recipe Wizard


Found this in the Subreddit, it’s another approach of automatic recipe generation. Typical results contain over 300 g of wheat flour, but it’s a start:

I’ve run the site’s JavaScript code through a beautifier and posted it here. It uses the Big M method for solving the linear optimization problem. @nickp, I don’t think it would be a big problem to integrate this algorithm in MakeSoylent. It seems much faster than the GA.

Looking at the Tumblr, Juan has been woking on this since May.


I saw it used sunflower oil. Isn’t that the kind of oil you want to stay away from? Is it possible to replace it with coconut oil or MCT oil?


Can you link to previous discussions on why you would want to avoid it?

Whatever the reason is, we should probably add it as a constraint to the optimizer (same for phytic acids).


Can’t find one but isn’t it heavy on Omega 6 when we should be going for Omega 3?


It is, but that doesn’t exclude the ingredient per se. The generated recipes indeed have a subotimal o-6:o-3 ratio of about 10:1, but that is because the optimizer only tries to satisfy the DRI. We can teach him to optimize the ratio as well.


I emailed Juan to see if he’d be willing to help.

I spent a while trying to do it myself but didn’t get too far. If anyone else wants to take a crack implementing the algorithm for makesoylent recipes, you can get the raw json for a recipe by going to a recipe page then adding ‘/json’ to the URL. eg:

The json contains both the nutrient profile and all the ingredients. Note that in each ingredient, the amount of an ingredient needs to be divided by the serving size in order to get the ‘per unit’ amount of a nutrient.