FODMAP Without Guessing

How GutCode's AI camera detects high-FODMAP foods from a single photo

5 min read

If you have ever been told to "try a low-FODMAP diet," you know what comes next: a sprawling list of foods to avoid, a shorter list of foods you can eat, and the sinking realisation that you are expected to memorise all of it while navigating restaurants, family dinners, and rushed weekday lunches. The mental load is enormous. Most people give up within a few weeks — not because the diet does not work, but because the logistics are exhausting.

GutCode was built to take that burden off your plate. Instead of memorising lists, you take a photo. The AI does the rest.

What are FODMAPs, exactly?

FODMAP stands for Fermentable Oligosaccharides, Disaccharides, Monosaccharides And Polyols. The concept was developed at Monash University in Melbourne, Australia, and has since become the gold standard dietary approach for managing Irritable Bowel Syndrome (IBS) symptoms worldwide.

FODMAPs are short-chain carbohydrates that certain guts struggle to absorb properly. When they reach the large intestine undigested, gut bacteria ferment them rapidly, producing gas. The resulting bloating, abdominal pain, cramping, and altered bowel habits are all too familiar to the estimated 10-15% of the global population living with IBS.

The tricky part: not every FODMAP affects every person. Your gut might handle lactose perfectly well but fall apart at the first hint of fructans. That individual variation is exactly why a blanket elimination list is such a blunt instrument — and why personalized detection matters so much.

The five FODMAP subgroups

FODMAPs are not a single substance. They break down into five distinct subgroups, each with its own set of common trigger foods:

• Fructans — Found in onion, garlic, wheat, rye, and many vegetables. Fructans are arguably the hardest FODMAPs to avoid because onion and garlic are the invisible backbone of most savoury cooking.
• GOS (Galacto-oligosaccharides) — Concentrated in legumes (lentils, chickpeas, kidney beans) and some nuts like cashews. A hummus platter or a bean chilli can be a minefield.
• Lactose — The sugar in dairy. Milk, soft cheeses, ice cream, and yoghurt are the usual suspects, though hard cheeses and butter are typically low enough to be tolerated.
• Excess Fructose — When a food contains more fructose than glucose, absorption can fail. Apples, pears, honey, mango, and agave syrup all fall into this category.
• Polyols — Sugar alcohols (sorbitol, mannitol, xylitol) found naturally in mushrooms, cauliflower, and stone fruits like peaches, plums, and cherries. They also appear in sugar-free gum and mints.

Understanding which subgroup is your personal trigger is the difference between a restrictive elimination diet and a precise, liveable one. But getting there requires data — and that is where most people hit a wall.

Why manual FODMAP tracking fails

In theory, you could look up every ingredient of every meal in a FODMAP database. In practice, this breaks down fast. Restaurant meals rarely list every ingredient. That Thai curry might contain garlic, onion, and honey in amounts you will never see on a menu. A store-bought pasta sauce could be loaded with onion powder and high-fructose corn syrup without any obvious visual cue. Even home-cooked meals become tedious to cross-reference when you are cooking for a family and juggling multiple ingredient lists.

The result? People either over-restrict (cutting out entire food groups "just in case") or under-track (logging meals inconsistently because the process is too slow). Neither approach gives you the data you need to identify your real triggers.

How GutCode's AI scan works

GutCode replaces the manual lookup with a single photo. Point your camera at your plate, tap scan, and the AI — powered by GPT-4o-mini vision — identifies every visible ingredient in the meal. Each identified ingredient is then cross-referenced against FODMAP risk categories and flagged with a clear warning if it falls into a high-FODMAP subgroup.

For processed and restaurant foods, the system goes a step further. If the AI detects a dish that typically contains hidden high-FODMAP ingredients — say, garlic powder in a store-bought marinara or onion in a pre-made stock — it flags those as suspected ingredients with confidence markers. You are not left guessing; the AI guesses conservatively on your behalf and tells you exactly what it suspects and why.

The entire process takes seconds. No scrolling through databases, no typing ingredient names, no second-guessing portion sizes. Scan, review the flags, eat.

From scan to personalized insight

A single FODMAP flag is useful. Hundreds of them, correlated with your symptom logs over weeks, are transformative. This is where GutCode's correlation engine — the same one behind the Gut Score — changes the game.

Every flagged FODMAP ingredient from your scans feeds into the engine alongside your symptom check-ins. Over time, the system does not just tell you "FODMAPs are bad" — it tells you which specific FODMAPs your gut reacts to. Maybe fructans are your clear trigger, but lactose causes you no issues at all. Maybe polyols only bother you at high doses. That level of precision is impossible with a generic elimination list, and it is exactly what the low-FODMAP diet was designed to achieve through its reintroduction phase — except GutCode gets there with less friction and more statistical rigour.

The three-zone journey

GutCode organizes your FODMAP ingredients (and all other flagged ingredients) into a clear three-zone system based on real data from your body:

• Watchlist — Ingredients under active observation. The engine has noticed them but needs more data points before drawing a conclusion. You will see a status like "Observing" here.
• Patterns — Confirmed triggers. These ingredients have shown a consistent statistical link to your symptoms across multiple meals. The badge progresses from "Possible Link" to "Consistent" as evidence accumulates.
• Safe Foods — Ingredients you have eaten repeatedly without any symptom correlation. These are your green-light foods — the ones you can eat confidently.

This is not guesswork or gut feeling. Every graduation from one zone to another requires multiple data points and a statistically meaningful pattern. The system is deliberately conservative — it would rather keep an ingredient on the Watchlist longer than promote a false trigger.

Beyond FODMAPs

FODMAPs are one of the most common triggers for digestive symptoms, but they are not the only ones. GutCode's AI scan also flags ingredients associated with other sensitivity categories — gluten, histamine, sulfites, and nightshades — so your trigger profile captures the full picture, not just the FODMAP slice.

This matters because real-world sensitivities rarely respect neat categories. You might react to both fructans (a FODMAP) and histamine-rich aged cheese (not a FODMAP issue at all). A tool that only tracks one category leaves blind spots. GutCode tracks them all in parallel, letting the correlation engine surface whatever pattern actually exists in your data — regardless of which dietary framework it belongs to.