You’re probably just gonna have to get better at guesstimating, (e.g. by comparing to similar pre-made options and their nutrition labels), or use an app for tracking nutrition that integrates with OpenFoodFacts and get a scale to weigh your ingredients. (or a similar database, though most use OpenFoodFacts even if they have their own, too)
I don’t really know of any other good ways to just take photos and get a good nutritional read, and pretty much any implementation would use “AI” to some degree, though probably more a dedicated machine learning model over an LLM, which would use more power and water, but the method of just weighing out each part of a meal and putting it in an app works pretty well.
Like, for me, I can scan the barcode of the tortillas I buy to import the nutrition facts into the (admittedly kind of janky) app I use (Waistline), then plop my plate on my scale, put in some ground beef, scan the barcode from the beef packaging, and then I can put in how many grams I have. Very accurate, but a little time consuming.
Not sure if that’s the kind of thing you’re looking for, though.
In that case, I’d say just find any food that’s just similar enough, and use that. It’s better to have a close-ish estimate than none at all.
For example, I had no clue what the nutrition would be like for the meatloaf I had the other day, so I just entered it as if it was pure ground beef and called it good enough.
You’re probably just gonna have to get better at guesstimating, (e.g. by comparing to similar pre-made options and their nutrition labels), or use an app for tracking nutrition that integrates with OpenFoodFacts and get a scale to weigh your ingredients. (or a similar database, though most use OpenFoodFacts even if they have their own, too)
I don’t really know of any other good ways to just take photos and get a good nutritional read, and pretty much any implementation would use “AI” to some degree, though probably more a dedicated machine learning model over an LLM, which would use more power and water, but the method of just weighing out each part of a meal and putting it in an app works pretty well.
Like, for me, I can scan the barcode of the tortillas I buy to import the nutrition facts into the (admittedly kind of janky) app I use (Waistline), then plop my plate on my scale, put in some ground beef, scan the barcode from the beef packaging, and then I can put in how many grams I have. Very accurate, but a little time consuming.
Not sure if that’s the kind of thing you’re looking for, though.
Actually, I am using waistline, but there are some food I can’t find and are hard to find nutritional values, and I am bad at guessing anything
In that case, I’d say just find any food that’s just similar enough, and use that. It’s better to have a close-ish estimate than none at all.
For example, I had no clue what the nutrition would be like for the meatloaf I had the other day, so I just entered it as if it was pure ground beef and called it good enough.
Yeah, true it is just that I kinda want to be perfectly accurate but yeah you are right