Food was the last piece

I have lived with diabetes for 17 years. I use a Libre 3, a pump, and a closed-loop system. My HbA1c is the best it has ever been, and most of the work runs in the background. Food is still the moment that makes me stop and think.

I love eating out. I love ordering whatever sounds good without doing mental math first. Takeaway nights, holidays, restaurants, or just being lazy after a long day all make food choices more spontaneous. I used to think I was great at carb counting and just accepted that this was the part I had to do myself. Even with all the automation in modern diabetes tech, food is still the one piece that feels manual. It is the last thing I need to actively manage.

Carb estimations are a huge help, but they are only the start. The real unlock is understanding how meals actually hit. Which choices keep me steady, which ones spike faster than I expect, and what small adjustments change the curve. It is not about changing what you love. It is about keeping your habits and doing them with more confidence. That is where trends and insights matter.

AI is part of that, but I do not want a black box answer. Models can be too confident. A core part of Bettie is being able to refine with chat until the estimate matches what you actually ate. The goal is to get to a real answer you can trust, not just a single guess.

Bettie is my attempt to close that loop. I want logging to be fast, and I want the learning to be automatic. The goal is to make food feel as seamless as the rest of the system. Snap a photo, get a reasonable estimate, and then see the impact without guessing.

I am building this for myself, but I hope it helps others who feel the same way. If food is the last piece of your puzzle too, I want you to feel the freedom that comes from understanding your own patterns.

If you want early access, you can join the beta here.