AI Nutrition Agent · Now on App Store

Decode every meal.

Delemma is an AI agent for nutrition. It turns your food, biometrics and goals into the next decision — automatically. No labels to read. No spreadsheets to keep.

Download on App Store Apply for TestFlight for beta review

304k

Training samples

3,200+

Dish categories

Nutrients tracked

Goal modes

Origin

It started with a single number.

Delemma began as one person debugging his own metabolism. The principles that worked turned out to be the principles that work for everyone.

In 2023, mid-thesis, jet-lagged, and chronically underslept, I was diagnosed with a runaway A1C of 14%. The clinic's plan was lifelong insulin.

I went home and started treating my body like a system to debug. Every meal logged. Every biometric tracked. Every recommendation cross-checked against the literature.

Sixteen months later A1C was 5.5% — back inside the normal range. The thing that worked wasn't a diet. It was a feedback loop: data in, decision out, signal back.

"What works for one person who's paying attention works for anyone — if the loop runs fast enough."

Rocky Li · Founder, Delemma

A1C · 21-month trace

Built for everyone in motion

Three goals. One agent.

Pick the goal that fits today's life. Delemma rewires food, biometrics and recommendations around it — and adapts as the goal changes.

Maintain & optimize

Stable life, modern diet — but tired, foggy, sleeping poorly. Delemma maps 49 nutrients against your real intake and surfaces the silent gaps: vitamin D, magnesium, omega-3, B-complex.

EnergySleepMoodSkin

Cut without breaking

Calorie deficits collapse without protein, magnesium, B-complex. Delemma separates "calorie gap" from "nutrient gap" so you keep your metabolism — not just your scale number.

Protein floorMacro splitsBMR guard

Fast with precision

16:8, 18:6, OMAD — the upside lives in the eating window. Delemma calculates exactly the protein, potassium, magnesium and sodium your window needs, and pre-stages electrolyte protocols.

Window mathElectrolytesInsulin sens.

Delemma Flux · the agent layer

An AI agent fed by
your real-time data.

Flux is the conversational core of Delemma. Every reply pulls live signal — biometrics, intake history, energy balance, your goal — and answers in the moment, not in the abstract.

Live biometrics. Heart rate, sleep, blood glucose, SpO₂ — pulled from Apple Health, Apple Watch, CGM.

Today's plate. The full breakdown of what you ate, and the 49-nutrient bank against your goal.

Seven-day trend. The nutrients you keep underconsuming, and the meal patterns that keep producing them.

Energy ledger. BMR + NEAT + activity + thermic effect of food, reconciled against the day's intake.

The Stack

Owned, end to end.

Two models, both ours: one reads the plate, one reads you. Both are trained, shipped and running in the Delemma app today — neither is a wrapper around somebody else's API, which is why your meals stay on infrastructure we own, and why the models keep getting better on our schedule instead of someone else's.

Delemma World Food Vision LM · live in the app

One photo. The whole plate, decoded.

Point the camera at dinner and it answers, end to end, what every other app leaves to you: what this is, how much of it there is, and the full 49-nutrient breakdown. Several dishes in one frame are itemised and totalled separately. When it isn't confident it offers candidates instead of bluffing — and if you point it at something that isn't food, it says so rather than inventing a meal. No camera to hand? Type what you ate and the same breakdown comes back.

Portion & weight estimation Multi-dish itemisation Refuses non-food images Candidates when unsure Text input works too Native EN & 中文 dish names

304K

Training samples

253K

Unique images

3,200+

Dish categories

4,287

Unique dish names

Source Samples Categories Share

Academic curated datasets

ChineseFoodNet127,072208

41.8%

Food-10168,220101

22.4%

UEC-Food25626,452256

8.7%

VFN9,30582

3.1%

Collected in-house

Global collection7,9101,366

2.6%

General collection7,175710

2.4%

Multi-dish plates8,018—

2.6%

Chain brands1,675187

0.6%

Restaurant picks1,350152

0.4%

Nutrition calibration3,355—

1.1%

Other small sets489—

0.2%

Capability layers · reuse existing images

Image-language28,792—

9.5%

Text-only layer14,463—

4.8%

Total304,2763,221

100%

Delemma Nutrient Reasoning LM · live in the app

Advice that knows who it is talking to.

Ask a general chatbot what to eat and you get the paragraph everyone gets. Ours was tuned for one field — and at run time, for exactly one person: your profile, your goal, what you have eaten today and how that sits against recent weeks, and the signals coming off your wrist. It holds your hard constraints across the whole conversation — allergens, pregnancy, the foods you have ruled out. Supplement advice arrives as specifications and strength of evidence, never a brand it wants you to buy. And it was trained to refuse invented nutrients and pseudo-vitamins rather than play along.

Anti-hallucination training Remembers your hard constraints Today vs recent trend No brands recommended Real-time voice No diagnosis, not medical advice

145K

Training samples

7,745

Validation samples

153K

Total samples

Source Samples Share

Authoritative food composition databases

USDA food database29,475

20.3%

Japan MEXT tables8,965

6.2%

China food tables5,941

4.1%

Nutrition knowledge & Q&A

Dish nutrition Q&A~25,500

17.5%

General nutrition Q&A11,897

8.2%

Supplement knowledge~4,500

3.1%

Diet-method reviews2,497

1.7%

Conversation & reasoning layers

Voice & reasoning layers~54,000

37.2%

Total145,322

100%

Ledger

Nutrients tracked · target vs intake

Live signal · wearables × meals

Your wrist and your plate, on one ledger.

01 · Signals in — synced in the background

Apple Health Apple Watch HealthKit CGM · via Apple Health (Dexcom / Libre / Lingo) Blood-pressure monitor · via Apple Health

02 · Reconciled — two sides of the same ledger

Burnedfour components

BMRMifflin-St Jeor

Exerciselogged workouts

NEATpriced to your body

TEFthermic effect of food

Eatenagainst your own targets

49 nutrients vs your own targets

= today's macro energy / micronutrients / hydration gaps

Dietitian team · free for every user

AI watches the numbers. A human steps in.

AI watches your numbers around the clock; a dietitian steps in when it actually matters. Every Delemma user has a dietitian team — at no cost.

They see the whole picture. Meals with photos, 49-nutrient attainment, ten vitals, and how consistently you've logged — structured data, not a vague note that you ate too much.

Alerts reach them first. Risk flags are queued to your dietitian by priority — you don't have to know to ask.

Feedback lands on one specific meal. You'll see \"About this meal\" in the app, attached to that plate's photo and its numbers.

In your language, on your channel. In-app message, SMS, email or push — sent in whichever language you use.

Reply however you like. Type it, or just say it out loud.

Nutrition management, not medical care — no diagnosis, and no substitute for professional medical advice.

Sustainable AI

85%+ less energy. Zero marginal cost.

We believe technological progress shouldn't cost the Earth. Delemma is engineered from the ground up as an eco-friendly AI agent. By optimizing our inference architecture, we cut energy consumption drastically compared to monolithic LLMs, allowing us to offer a sustainable free tier for everyone.

Small & Focused

Specialized SLMs

General-purpose giants burn a staggering amount of compute to answer a question about lunch. Ours are small, nutrition-specialised models — a fraction of the footprint, with the expertise concentrated exactly where it matters.

Local First

Edge Inference

Photo parsing runs on lightweight nodes close to you; the heavier reasoning only wakes when a question genuinely needs it. Most of what you ask never troubles a large model at all.

Zero Waste

Incremental Compute

We cache nutrition embeddings. When your intake doesn't fundamentally change, the AI doesn't re-compute. BMR/TDEE math runs purely on-device with zero network energy.

Roadmap

Now live on the App Store.

Delemma launched on the App Store on Jun 21, 2026. Both models — vision and reasoning — are trained, shipped and serving real users; TestFlight beta continues to run in parallel.

Delemma App for iOS 100%

Live on the App Store — Jun 21, 2026. TestFlight beta also available.

Delemma Nutrient Reasoning LM 100%

Trained, shipped, live in the app — personalising every meal note and daily read, and still evolving with each release.

Delemma World Food Vision LM 100%

Trained, shipped, live in the app — every photo runs through it, end to end into the 49-nutrient ledger, and it keeps evolving with each release.

Nov 2025 Project kickoff

Apr 27, 2026 Final integration · 85%

Jun 15, 2026 TestFlight live

Jun 21, 2026 App Store · live

Product

In your pocket.

Four screens, one loop. Log a meal, see the gap, get the move, close the gap.

Daily home dashboard — 01 · HOME Daily home Today's calorie balance, meal log and supplement check on a single glance.

02 · STACK Supplement intelligence Your stack, scanned, evaluated against today's gaps and goal — keep, swap or drop.

One-tap food recognition — 03 · CAPTURE One-tap recognition Photograph the plate. Get the full breakdown — calories, macros, micros, GI.

04 · LEDGER 49-nutrient ledger Every nutrient, percent of target, delta to close. Sorted by what matters today.

From the early users

It clicks fast.

A handful of testers running the loop for a few months. Different bodies, different goals, same pattern.

Sarah Mitchell

Distance runner · 34

Daniel Klein

Software engineer · 16:8 IF · 29

Akira Tanaka

Tokyo · cut 12 kg, kept it · 41

Maya Reyes

Fitness instructor · mom of two · 38

Ben Lawson

Biohacker · CGM + Apple Watch · 45

Liu Wei

Beijing · product manager · 33

The name

Why "Delemma"?

Every meal looks like a dilemma. Add the prefix and it stops being one — the way you de-bug code, you de-lemma a decision.

de-

Remove the burden. As in de-bug, de-fog, de-fuse — undo the thing in front of you.

lemma

A useful step. In math, the lemma you prove on the way to the theorem. Each data point is one — read enough, the answer reveals itself.

.ai

Where the agent lives. A homophone for Llama — a nod to the open-source lineage this whole field grew out of. Intelligence as the default, not a feature.

Partnership

Building this with the right partners.

Delemma is live, the whole stack is ours, and it keeps evolving. We're open on two fronts.

Investment & strategic partnership

Open to the conversation.

Two models we trained ourselves, the 49-nutrient ledger, the live biometric pipeline, and the reasoning that ties them together — all built and run in-house, and all in users' hands today. If you invest in health or in AI infrastructure and this is your thesis, we're open to investment and strategic partnership conversations. An email is the whole first step.

Self-owned model stack Live in production EN · 简体 · 繁體

Clinics & dietitian practices

Bring your own clients onto Delemma.

Run a nutrition clinic, a hospital nutrition department or a dietitian practice? You can manage your own clients on Delemma. What you get is a console you can actually work in rather than a dashboard to look at: each client's meals with photos, their 49-nutrient attainment, ten vitals, how consistently they've logged, and risk flags already sorted by priority. Leave feedback on one specific meal and it reaches the client attached to that plate. Messages go out over in-app, SMS, email or push — automatically in each client's own language.

Client console Priority-sorted alerts Per-meal feedback Auto-localised outreach

Investment enquiries and clinic onboarding both start with an email.

Mission

Nutrition is a daily decision. Make it disappear.

Most health apps stop at calorie counting. The rest demand you read the literature. Delemma does the reading, runs the loop, and hands back a single next move — backed by the real numbers from your day.

From dilemma to Delemma.