Building a meal plan for a patient isn't quick. You're balancing nutrient targets, allergens, medical conditions, cultural preferences, and patient feedback; all at once, all by hand. Most dietitians were doing it in spreadsheets, rebuilding the same structure for every patient, every session.
The software that existed either generated generic plans nobody could actually use, or locked things down so tightly that clinicians rejected it entirely. There was no middle ground between "too simple" and "too rigid."
We built the middle ground.
Compumeal is a clinic platform with AI in three specific places, just where it makes a measurable difference.
The meal plan generator starts with a linear-programming solver that hits the patient's exact nutrient targets mathematically. Then a fast, cheap AI model scores the result for realism, does this actually look like food a person would eat? Only when the score falls below a threshold does a more powerful model step in to make targeted swaps. That conditional logic means AI is only used when it's needed.
The in-editor assistant sits next to the meal plan and lets the dietitian make changes in plain language. "Swap the chicken for something plant-based" or "reduce sodium in the Thursday dinner." Every suggestion comes back as an approval card; the dietitian stays in control, the AI does the heavy lifting.
A third AI layer handles bulk food classification; automatically tagging imported foods with meal type and dietary compatibility so the database stays usable without manual upkeep.
The cost economics had to work.
This was the constraint that shaped everything. If the AI cost too much per plan, the product couldn't be sold at a price a small clinic could afford.
The answer was a tiered architecture. A cheap classifier decides how complex each request is. A cheap evaluator scores the plan. The expensive model only runs when the cheap one says it's not good enough. That conditional approach brings the AI cost per meal plan to under 2¢; compared to 30–50¢ if you just called the most powerful model every time.
For a small clinic generating 50 plans a month, the total AI cost is around $14. For a large organisation generating 1,000, it's around $370. The economics work at every scale.
Built once, rebuilt better.
Compumeal was originally built in 2024. We rebuilt it from scratch in early 2026; new stack, new AI architecture, better cost structure. The rebuild took 12 weeks. The product has been live with real signup traffic since March 2026.
“We built our own production AI product and made the cost economics work before we shipped a single plan."
- Meal plan generation from patient nutrient targets; mathematically optimised, AI-refined
- AI cost under 2¢ per plan; through conditional model usage, not cheaper models
- In-editor AI assistant with dietitian-controlled approval cards
- Rebuilt from scratch in 12 weeks with a production-tested cost architecture
- Live with real users; domain sign-up protection deployed March 2026
- $14/month AI cost for a small clinic at typical usage
