I use ChatGPT (4o model) to track my daily macros, specifically calories, protein, fat, and carbs. I have a prompt containing my macro targets and a list of the foods I typically eat or have at home. Throughout the day, I report my meals, for example five large eggs, two slices of sprouted wheat toast, and a cup of spinach with a pat of butter. It calculates the macros for each meal and keeps a running macros total so I can see how I’m tracking. It’s so easy—I don’t even type, I use (the overall excellent) ChatGPT’s TTS function in the iOS app.

If I eat something unexpected, I have it recalculate what I need for the rest of the day. If I were planning on 8 ounces of chicken for dinner but had an unplanned snack, it might tell me that 6 ounces will suffice now, and I now have room for more fat, so I could add some cheese. I find it quite accurate in this.

Where it loses accuracy is over time. Even when focused on just a single day, the cumulative totals can be wrong when the context window gets too large. I catch this by mentally adding up the numbers it spits out. If the numbers seem off, I open a new chat. Also, it’s not reliable for looking back over multiple days or trying to track long-term patterns.

When I take a photo of a meal (like when eating out), it performs very well. I ask it to list the contents of the plate, then I correct any assumptions. Once corrected, the resulting macros are usually within 5% accuracy. I also tell it the quality of the ingredients and restaurant overall, which can help a bit.