Why I Built Avocavo — A Nutrition API Powered by USDA FoodData Central for Developers

A few months ago, I was working on a project that needed to parse recipes and return nutrition data.

Being an engineer (and maybe going a little overboard), I ended up spending way too much time on the nutrition calculation part — making sure it was accurate, fast, and worked for every oddball input I could throw at it.

At some point I realized:

If I'm going through all this pain, a lot of other developers probably are too.

I didn't want to just wrap GPT — it's not something I can fully trust for accuracy, and I can't control what people feed into it. But AI is still a useful tool for other parts of the workflow, so I've used it where it makes sense.

About me

I have an electrical engineering degree and spent my career as a nuclear systems engineer and substation engineer before moving into software and algorithm development. I've built algorithms for hedge funds, designed high-reliability engineering systems, and been tinkering with software and hardware since I was 8 years old (I'm almost 40 now!).

This is my first time building something specifically to scale for anyone to use, rather than just for one company or internal project.

Biggest challenge

Making something that's both accurate and fast is hard.
Making it accurate and fast while scaling to a lot of users is even harder.

I've spent a lot of time optimizing the parsing and matching so it can return results powered by USDA FoodData Central in ~300–400 ms, whether it's a single ingredient or a full recipe.

What excites me

I can't wait to see what people build with this. I'm already gathering tons of recipe data so it's always improving, and I'd love to hear what other features developers would find useful.

If you want to try it

https://nutrition.avocavo.app

You can throw any ingredient or recipe at it and get clean nutrition data powered by USDA FoodData Central back in JSON. Free tier available — no card needed.