Internal · positioning input · August 2026

Market & competitor analysis

Where emealia sits in the nutrition-app market after the 2026 consolidation, and what the new feature additions — band contributions, multi-meal logging, manual weight, mood and activity tracking — are actually worth competitively. Successor to the 2026-07-07 analysis, which stays authoritative for the science base, the German willingness-to-pay data and the COGS model.

00Bottom line 01What changed since July 02The field, August 2026 03Feature matrix 04The new features, assessed 05Deterministic tip library 06Differentiation opportunities 07Advantages & exposures 08Recommendations 09Watchlist 10Caveats

00 · Bottom line

Four things changed in the four weeks since the last analysis, and three of them cut against emealia's original wedge.

MyFitnessPal closed and absorbed Cal AI, so the incumbent now owns low-friction AI input. MyFitnessPal then moved more of its free tier behind the paywall and caught a class action for it. And Apple is building native nutrition logging plus a paid AI health coach into the platform itself.

"Easy AI logging" is no longer a differentiator. It is now table stakes owned by the two largest distributors in the category. What survived intact is the frame — honest ranges instead of point numbers, colored bands instead of a budget, no red, no streaks, no verdict. Nobody credible is competing there, and the two 2026 scandals hand emealia its clearest marketing arguments for free.

Feature	Verdict	Why
Band contributions	Strategic	The only genuinely novel thing in the list. It closes the biggest logical hole in the band concept (a band says where, never why), adds perceived intelligence at zero AI cost, and its output is the best possible input for the weak suggestions layer. Also the highest ED-safety risk surface in the product.
Multi-meal logging	Strategic · underrated	Nobody in the field does true batch free-text entry. It directly answers the category's best-documented complaint (entry fatigue) and rescues the evening catch-up user — the exact moment every tracker loses people.
Mood + activity (manual)	Necessary, not differentiating	Standard in the guilt-free-journal segment (Eating Buddy, Munch) and present in the coaching segment (Lifesum, Noom). The differentiation is not having it — it is refusing to mine it for behavioral causality the way Noom does.
Weight tracking (manual)	Defensive · keep it small	Universal table stakes. Its only strategic job is removing an objection. Do not build a goal-weight progress bar; that is a number-as-verdict, the thing the product exists to avoid.
Suggestions	The real gap	Competitors are shipping AI coaches fast (MFP, Simple, Noom, soon Apple). emealia cannot win on coaching breadth. It can win on coaching restraint — and contributions give it the cheapest, most concrete raw material available.

01 · What changed since 2026-07-07

1.1 MyFitnessPal closed the Cal AI acquisition and is winding the app down

The deal signed in December 2025 was announced 2 March 2026; Cal AI's standalone app is being wound down and its photo-AI recognition folded into MFP Premium (TechCrunch, The Nutrition Magazine). Cal AI had passed 15 M downloads and >$30 M ARR in under two years.

Consequence. The July analysis treated Cal AI as the benchmark for a low-friction upstart beating incumbents. That window closed by acquisition, not by competition. The incumbent with the largest food database also has the best consumer AI-input funnel. emealia's wedge can no longer be "logging is easier here" — MFP's meal scan is at least as easy. It has to be "logging is easier here and the result isn't a verdict."

1.2 The May 2026 paywall expansion and the class action

In early May 2026 MFP moved scan-a-meal photo logging, recipe-URL import and macro-by-meal goal tracking into Premium (The Nutrition Magazine). A consumer class-action complaint followed the same month, alleging the app's store metadata still markets it as "free" (Consumer Tech Wire). Tiers now: Free · Premium $79.99/yr or $19.99/mo · Premium+ $99.99/yr (breakdown).

Two consequences. (a) The strongest possible external validation of non-negotiable #3 — one plan, everything in it, forever. Quote it, don't paraphrase it, in comparison SEO. (b) Note which feature went behind the wall: per-meal macro targets. The incumbent has decided meal-level nutrient attribution is worth paying for. emealia's contribution view is adjacent to that — and free.

1.3 Apple is building the category into the platform

The Health app redesign (iOS 26.4, spring 2026) adds native nutrition logging — meals, calories, macros, directly in Health (9to5Mac, Wareable) — and a paid Health+ tier with an AI coach that suggests nutrition adjustments from Watch and iPhone data (Wareable).

Consequence. Basic logging and generic AI tips are being commoditized to zero marginal price for a very large installed base. Everything emealia does that Apple will also do is worth nothing strategically; everything Apple structurally won't do — non-numeric bands, no verdict, an opinionated wellbeing posture — is worth more than it was a month ago. Apple will ship numbers, rings and goal completion. It will not ship "no numbers on the main view".

Being a PWA is a genuine hedge (no App Store cut, no review, no gatekeeper) and a genuine cost (no store discovery, and — see §4.4 — no HealthKit access at all).

1.4 EU AI Act transparency obligations are now live

Article 50 transparency obligations became enforceable 2 August 2026 (Cooley, European Commission): people must be clearly informed when they interact directly with an AI system, at the appropriate point in the journey, with accessible rather than buried disclosures.

Consequence. emealia is unusually well positioned — it already labels every estimate as an estimate and shows ranges rather than false precision, which is the substance of the obligation, not just the notice. But this needs an explicit compliance line item alongside the GDPR epic (AI-interaction disclosure at onboarding and at the meal-input surface), and it is a strong DE marketing argument next to the existing DSGVO one.

1.5 Yazio raised prices again

Yazio's German list price is now €83.90/yr with rotating discount framing, and criticism has hardened around inconsistent pricing by entry point and a trial only available on the web, not in-app (Check-App). Yazio still has no competitive AI photo logging.

The German incumbent is now priced at roughly 2.1× emealia's annual price with a worse input method and an actively resented pricing experience. This is the best head-to-head comparison emealia has in its primary market — and it should be the first comparison landing page.

1.6 GLP-1 has become the category's centre of gravity

MFP shipped GLP-1 medication tooling (MobiHealthNews); Noom runs a GLP-1 Companion; a whole dedicated segment exists (MeAgain, Shotsy, Glapp, Pep, Phaze).

Consequence. A large, high-willingness-to-pay slice of the market is being pulled toward medication-adjacent products. emealia should not follow — it is a medical claim surface, it conflicts with the general-wellness positioning still pending legal sign-off (H10), and it is a market where capital wins. The useful read is the opposite one: as competitors chase GLP-1 users, the non-medicated "I just want to eat better" majority is being under-served. That is precisely emealia's persona set.

02 · The field, August 2026

Seven segments. emealia occupies the boundary between the precision trackers, the AI photo-first wave and the guilt-free journals — the unoccupied middle the July analysis identified, still unoccupied.

A · Mainstream freemium trackers

App	Price 2026	Input	Core weakness
MyFitnessPal	Free · $79.99/yr · $99.99/yr Premium+	Search, barcode, voice, meal scan (Cal AI), AI coach	Crowdsourced DB error rate; paywall creep; May 2026 class action; most expensive mainstream option
Yazio DE	~€83.90/yr list, heavily discounted	Search, barcode	Macros behind paywall; price confusion; no competitive AI input; trial not available in-app
Lifesum	~€8–10/mo · ~$45/yr Premium	Search, barcode, AI scan	Accuracy and stability complaints since the 2024/25 AI pivot
Lose It!	~$39.99/yr	Search, barcode, Snap It	~10% DB error rate; per-meal breakdowns Premium-only
MyNetDiary	Freemium	Search, barcode, voice, autolog of routine meals	Conventional numeric framing
FatSecret	Free + ads	Search, barcode	Ad-supported; dated UX

B · Precision / pro trackers

App	Price 2026	Position
Cronometer	Free w/ ads · Gold $49.99/yr	84 nutrients, USDA/NCCDB-verified, <5% macro error. Steep learning curve. The accuracy benchmark.
MacroFactor	$11.99/mo · $71.99/yr · $89.99/yr bundle	No free tier. Adaptive TDEE from weight trend is the USP. Best-in-class algorithm, hard paywall, generous refunds.

C · AI photo-first

Cal AI is being wound into MFP. The rest is a long tail of 2026 entrants — PlateLens, Welling, SnapCalorie, Foodvisor and dozens of clones — competing on recognition speed and marketing spend.

Caveat that matters. The "independent benchmarks" ranking these apps are almost all published by competing apps' own content marketing (Nutrola, PlateLens, Welling, NutriScan, Amy Food Journal all sell apps). Treat every accuracy claim in that ecosystem as advertising. Peer-reviewed and neutral work still puts AI food estimation at roughly 15–25% mean absolute error, with a 2025 restaurant-condition RCT finding 86% of dishes correctly identified but only 68% accurately reported end-to-end once portion entry was included.

D · Behavioral coaching / GLP-1

App	Price 2026	Position
Noom	~$24.99/mo · $59–199 programs	CBT lessons, mood + meal + weight + sleep dashboards, GLP-1 Companion. Reputation damage: 1.5-star consumer aggregates, BBB "D", >1,200 cancellation complaints.
Simple	$49.99–59.99/yr	Coach Avo (AI), Avo Vision photo feedback
Calibrate · MeAgain · Shotsy · Glapp · Pep · Phaze	varies	Medication-centric: dose schedules, side effects, shot day

E · Guilt-free journals — emealia's philosophical neighbours

App	Price	What they track
Ate / AteMate	Free + small premium	Photo journal, "reflective health journal, not a diet app"
Munch	Free	Meals, hunger levels, emotional triggers; explicit intuitive-eating / binge-eating framing
Eating Buddy	Free + premium	Meals, hunger, fullness, mood, workouts, sleep, drinks

This segment already does mood and context tracking as standard. What it does not do is nutrients. That gap — context and nutrition, neither weaponized — is emealia's actual position.

F · Product scanners

Yuka — 76 M+ users, free scan + score, barcode-only, criticized by nutrition professionals for oversimplification (penalizes nut butter for energy density; the 10% organic bonus is not evidence-based). Proves enormous demand for a simple "is this good for me" signal. Not a direct competitor; a demand indicator.

G · The platform new in 2026

Apple Health / Health+ — native nutrition logging in iOS 26.4, paid AI coach to follow. Free, pre-installed, default. The most important new entry in this table and the one nobody can out-distribute.

03 · Feature matrix

● full ◐ partial / paid-only ○ absent ● uncontested emealia row

	emealia	MFP	Yazio	Cronometer	MacroFactor	Lifesum	Noom	Cal AI¹	Ate / Munch / EB	Apple Health+²
Free-text AI logging (primary)	●	◐	○	◐	◐	◐	○	◐	○	◐
Photo logging	○ Plus	◐	○	●	○	●	○	●	●	◐
Barcode / food database	○ by design	●	●	●	●	●	●	◐	○	◐
Ranges instead of point numbers	●	○	○	○	○	○	○	○	—	○
Explicit confidence level	●	○	○	○	○	○	○	○	—	○
No numbers on the daily view	●	○	○	○	○	○	○	○	●	○
No red / no failure state	●	○	○	○	◐	○	○	○	●	○
No streaks / no gamification	●	○	○	◐	●	○	○	○	●	○
Per-meal contribution to nutrients	●	◐	○	◐	○	○	○	○	○	○
Multi-meal logging in one entry	●	○	○	○	○	◐	○	◐	○	○
Manual weight tracking	●	●	●	●	●	●	●	●	◐	●
Adaptive target from weight trend	○ deliberate	○	○	○	● USP	○	◐	○	○	◐
Mood tracking	●	○	○	○	○	●	●	○	●	◐
Activity tracking	● manual	●	●	●	●	●	●	◐	●	●
Fitness-app / wearable sync	○ PWA	●	●	●	●	●	●	◐	◐	●
Micronutrients	● 32	◐	◐	● 84	◐	◐	○	◐	○	◐
AI coaching / suggestions	◐ weak	●	○	○	●	◐	●	○	○	●
GLP-1 support	○ deliberate	●	○	●	○	◐	●	○	○	◐
Offline logging	●	◐	◐	◐	◐	◐	○	◐	●	●
No ads · no data sale · EU residency	●	○	○	◐	●	○	○	◐	◐	◐
Everything in one price	●	○	○	○	●	○	○	◐	◐	○
Annual price	€39.99 / $44.99	$79.99–99.99	~€83.90	$49.99	$71.99	~$45	~$319	$29.99	free–small	tbd

¹ Cal AI as of the wind-down announcement; capabilities migrating into MFP Premium. ² Health+ specifications are pre-release reporting, not shipped product — directional only.

What the matrix says. emealia is not competitive on breadth, integrations, database depth or coaching maturity, and will not become so. It is uncontested on eight rows — ranges, confidence, no-numbers, no-red, no-streaks, free retrospective contribution, batch text logging, one-price-everything — and those eight rows are the entire product thesis.

The instruction that falls out of this: never trade a non-negotiable row for a breadth row. Every row emealia loses is a row where it was never going to win.

04 · The new features, assessed

4.1 Band contributions strongest addition

What it is. Tapping any nutrient band, on Today and in All-details, reveals which of the day's meals fed that nutrient and by how much, relative to each other.

Market precedent. Genuinely thin. What exists is prescriptive and numeric: MFP's per-meal macro targets (now Premium), Lose It! Premium's per-meal breakdowns, Nutrola's per-meal goal split, RP Diet's prescribed per-meal protein. All of these split a target forward — "you should get 40 g at lunch" — and all produce a compliance judgment per meal. Cronometer lets you derive attribution by reading a table. Nobody surfaces retrospective, relative, non-numeric attribution as a first-class explanation.

Why it matters more than it looks

It closes the band concept's one logical hole

A band tells the user where the day sits. It has never told them why. An explanation-free signal is functionally a verdict — the exact thing the product exists to avoid. Contribution turns the band into an observation with evidence: "glass box" at the day level, not only at the meal level.

Perceived intelligence at zero variable cost

Deterministic arithmetic over data already in daily_summaries. No AI call, no tokens, no hallucination surface, no UsageMeter reservation. Features that raise perceived intelligence while lowering COGS are rare.

It is the missing input for suggestions

The contribution ranking is the reasoning a coaching model would otherwise derive from raw meal history. See §4.6.

It only reads well in emealia's frame

"Lunch was 43% of your protein" inside MyFitnessPal is an invitation to arithmetic and a per-meal grade. Inside a range-based, no-numbers product it is just a fact about the day. The feature fits better here than at a competitor — the definition of a hard-to-copy feature.

Risks — the highest-risk surface in the product

Attribution is one word away from blame. "71% of your fat came from the pesto pasta" is an accusation if copy or ordering implies fault. Rank by magnitude but never label a rank ("largest contributor", never "worst"); use the meal's own name, never a judgment adjective; keep category-color = which nutrient, never good/bad; never pair the list with an above-zone warning. Non-negotiables #1 and #6 must govern this list explicitly, in the acceptance criteria.
The percentages will not add up, and users will notice. Per §08 of the use-cases doc, pending and low-confidence meals contribute zero to the day's ranges. If they are silently excluded, shares look wrong. Show them as an explicit calm row ("still estimating — not counted yet"), consistent with non-negotiable #6.
The math must be honest about which quantity is shared. Day ranges aggregate by root-sum-square, so uncertainties do not add. A share can only sensibly be computed on midpoints and should read as approximate ("about a third"), not as a precise percentage. A two-decimal share of an estimate with ±25% error is exactly the false precision the product rejects.
Do not let it become a per-meal grade. The unit of judgment in emealia is the day. Consider capping the list (top contributors + "the rest"), matching the existing top-5 dominating-ingredient pattern.

Recommendation. Treat contributions as a first-class product surface, not a detail view. Specify it in the band-state design language (3-design.md + design system), not only in a story. Add it to the G1 human gate: if testers read the contribution list as blame, the copy is wrong and shipping it as-is would undo the band work.

4.2 Multi-meal logging underrated

What it is. One free-text entry describing several meals, producing several logged meals.

Market precedent. None that matters. The field's answer to entry fatigue is templates and repetition — MyNetDiary autologs routine meals, Lifesum saves meal combos, Cal AI re-logs remembered meals, everyone has favourites. All of these speed up repeating yesterday. None lets you describe an unrepeated day in one go. Cal AI's founding story is that manual logging is tedious; the industry answered with faster single entries, not fewer entries.

It rescues the exact moment trackers lose users. The evening catch-up — "I haven't logged since breakfast, so why bother" — is the canonical abandonment point, and it is a failure state by construction. Batch entry turns it into one sentence. That is non-negotiable #6 applied to the logging flow itself, and it plausibly moves D7 more than anything else on this list.
It is cheaper per meal. One prompt producing three meals costs roughly one call's overhead instead of three. In a 30-credit trial the user feels that as generosity.
It is only possible because of the free-text architecture. A barcode/database product structurally cannot offer it. A defensible consequence of the core bet.

Open decisions

Decision	Recommendation
Credit accounting	One credit per AI call — one for the batch, not one per extracted meal. Rewards the wanted behavior, makes the trial feel generous, trivially explainable. The anti-farming objection is better handled by the existing per-user daily cap plus a hard max-meals-per-batch limit than by credit arithmetic.
Accuracy degradation	Splitting and estimating in one prompt is harder, and boundary errors are likely ("…and a coffee" — separate meal or part of breakfast?). A batch should produce N drafts on one review screen where the user can merge, split, retype or drop before saving. Never auto-save a batch.
KPI hygiene	"Meals per active user per day" becomes partly a function of this feature rather than of engagement. Keep the north star at weekly active loggers with ≥4 logging days and treat meals/day as a diagnostic.

Recommendation. Market it, don't just ship it. "Log your whole day in one sentence" is sharper and more copyable than "log a meal in one sentence", it is verifiably unique in the field right now, and it carries "emealia meets you where you are" better than anything currently on the landing page.

4.3 Mood tracking (manual) necessary, not differentiating

Market precedent. Common. Lifesum ships a daily mood tracker; Noom tracks mood alongside meals, weight, water and sleep and explicitly markets AI insight into "emotional triggers"; Eating Buddy tracks mood, workouts, sleep and drinks; Munch logs hunger and emotional triggers. Mood is standard in both the journal and the coaching segment. Nobody will be impressed that emealia has it.

Where the differentiation actually is. Every competitor that tracks mood uses it to explain the user to themselves — "you eat more when stressed". That is Noom's core mechanic and the most guilt-prone surface in the category, because a causal claim about your own weakness is much harder to shrug off than a number. emealia's opportunity is the restraint: log it, show it as context, never assert causality.

Compliance note. Mood data plausibly falls under GDPR Art. 9 on the same logic as body data. It belongs in the existing consent ledger with its own consent type, independently erasable, and excluded from any AI context not covered by that consent. Launch-blocker-adjacent, not a nice-to-have.

Recommendation. Keep it as calm day context in History and Statistics, exactly as daily_context_tags already models it. Do not build mood↔nutrient correlation. If it ever ships, put it behind explicit opt-in and phrase it as an observation the user can reject, never as a finding.

4.4 Activity tracking and the fitness-app question

Manual activity is fine, cheap and already modelled (activity_context plus the movement×sport PAL grid in onboarding). No issues.

Integration is a bigger decision than it appears, for two independent reasons.

1 · It is largely unavailable to a PWA

HealthKit is native-only; there is no web API. Google Health Connect is likewise an Android platform API. A PWA cannot read Apple Health or Health Connect, full stop. Shipping "Apple Health sync" would require a native wrapper — reintroducing App Store review, the 15–30% cut and store-listing exposure the strategy deliberately avoids. What is feasible over the web: Garmin Connect, Fitbit, Withings, Strava, Polar OAuth cloud APIs, server-to-server, no native code.

2 · Burned calories reintroduce the worst mechanic in the category

"You earned 340 extra calories" is the food-as-currency loop that makes MyFitnessPal painful for exactly the users emealia is built for, and it is well represented in the disordered-eating literature the strategy already cites. Crediting activity against the energy band would be a non-negotiable violation dressed as a feature.

Recommendation. Position activity as context that shifts the ideal zone slightly, never as a balance to be spent. If integrations are built, build Garmin/Fitbit/Withings/Strava over OAuth and use them only to refine the PAL factor. And say publicly why Apple Health sync is absent — "we're a web app, so your health data never leaves the browser for a platform account" is a better story than an apology.

4.5 Weight tracking (manual) keep it small

Market precedent. Universal. MacroFactor's whole product is doing something with the weight trend; Noom, MFP, Yazio and Lifesum all chart weight with a goal line.

Assessment. Purely defensive. Its job is to stop the weight-aware persona bouncing, and it already has the right posture in strategy: available from Profile and statistics, never the north star, never the only definition of progress, collapsed when the selected goals are not weight-related.

The one thing not to build: a goal-weight progress bar or "X kg to go" countdown. That is a number as a verdict with a failure state built in — the clearest contradiction of non-negotiables #1, #2 and #6 that could plausibly be argued into the product on grounds of "everyone has it". A smoothed trend line with no target line is the range-compatible equivalent.

On MacroFactor-style adaptive targets: genuinely good, and range-compatible in principle ("your energy zone has shifted a little"). But it requires consistent weighing, which conflicts with both the low-friction promise and the ED-safety posture. Recommend explicitly not building it — and recording that as a decision rather than an omission.

4.6 Suggestions the real gap

This is the weakest surface at the worst time. MFP has an AI nutrition coach; Noom has a lesson engine and emotional-trigger analysis; Simple has Coach Avo; MacroFactor has algorithmic coaching that actually adapts; Apple Health+ is bringing an AI coach with the platform's data advantage. Coaching is becoming table stakes and emealia cannot win on breadth.

Restraint as a feature

Silence by default, one forward-looking tip, never "eat less", no lesson treadmill, no "you haven't logged in 3 days" nagging. Against Noom's documented reputation damage and MFP's chatbot, "we will not talk at you" is a real, marketable position — and it is cheap, which matters at ~€0.08/user/month.

Concreteness via contributions

Feeding the model the top contributors plus the shortfall nutrient, instead of a bounded meal log, means fewer tokens, less personal data in the prompt, and a more specific tip. A tip anchored to a named contributor — "the lentil soup carried most of your fibre today; a similar bowl tomorrow would do the same" — is additive, derived rather than invented, and structurally incapable of shaming, because it praises something that happened.

Recommendation. Do not scope suggestions as a coach. Scope it as "one sentence that names something you already did and extends it", and build the contribution → suggestion pipeline as the primary path. Achievable, cheap, on-voice, and differentiated in a way a bigger coaching engine cannot copy without changing its own frame.

05 · The deterministic tip library — suggestions without AI

The premise: suggestion selection is a versioned rule table returning an opaque key; the AI, if used at all, only phrases it.

This mirrors classify_band_state exactly — the evaluator knows the rule shape, never what a key means — and it buys four things a model cannot: it is free, unit-testable, auditable (you can prove which rule fired, which matters under AI Act Art. 50), and structurally incapable of inventing a health claim.

5.1 What can be triggered from

Everything below is computable from data emealia already stores. No new schema, no new consent.

Source	Available signal
`daily_summaries` × `user_targets`	Per-nutrient day ranges against the ideal zone; band states; the day level
Contributions	Per-meal share of each nutrient — the newest and richest signal in the set
`meals`	Timestamps, `meal_type`, the merged `ingredients` list (`text_en`, `amount`, `dominating`), portion size
`meal_memory`	Confirmed repeat dishes in the user's own words, with confirmation counts
Logging pattern	Which days have logs, how many, first-meal time spread
`daily_context_tags` · `activity_context`	Mood, hunger, energy, rough activity — modifiers only, never triggers
Profile	Goals list + free text, cooking level, nutrition-knowledge level, dietary preferences, locale

5.2 Rules every tip must obey

Additive or substitutive, never subtractive. No tip may contain less, avoid, cut, reduce or stop. Where the honest advice is reduction (sodium, sweet drinks), it must be reframed as an addition or a one-for-one swap — never as removal.
Forward-looking. About tomorrow or the next meal, never a verdict on today.
Anchored in something real from their own data — a dish they ate, a pattern they have. Never generic.
No medical claim. No disease, deficiency, diagnosis or supplement recommendation. Any nutrient whose honest advice is medical is excluded from the library entirely (see §5.4).
Never triggered by absence. Not logging is not a signal, per silence-is-default.
Never a comparison — not to other users, not to a past self, not to a "should".
One per day maximum, not before 10:00, cutoff 23:00, minimum 3 logging days all-time. No deficit-shaped tip for the same nutrient two days running.
If number-hiding is on, the copy contains no numbers.

5.3 The library

Evidence bar: pooled human outcome data (meta-analysis / RCT) or explicit guideline consensus. Copy is EN draft, in emealia's register — DE follows once the set is signed off.

Key	Fires when	Tip (EN draft)	Evidence
`fibre_legume`	Fibre below zone on ≥3 of the last 7 logged days and no legume in those days' ingredients	"Lentils, beans or chickpeas are about the easiest fibre there is — a handful in tomorrow's lunch carries a surprising amount of the day."	Fibre 25–29 g/d → 15–30% lower all-cause and CV mortality (Reynolds 2019, Lancet); legume dose–response
`fibre_wholegrain`	Carbs at/above zone and fibre below zone, ≥3 of 7 days	"Same bread, wholegrain version. It's the smallest swap on this list and one of the best studied."	Per 90 g/d whole grain: ~17% lower all-cause mortality (Aune 2016, BMJ)
`plant_diversity_nudge`	Distinct plant ingredients over a rolling 7 days < 20	"You've had 14 different plants this week. Herbs, seeds and spices all count — a couple more kinds is one of the better-evidenced small moves."	Plant-type diversity predicted microbiome diversity more strongly than diet label (American Gut Project, McDonald 2018)
`plant_diversity_praise`	≥30 distinct plants in 7 days	"Thirty different plants this week. That's the number the research keeps landing on — nothing to change."	as above
`protein_breakfast`	Protein below zone and breakfast's protein contribution < 15% on ≥3 days	"Most of your protein arrives later in the day. A bit of it at breakfast tends to make the morning steadier."	Even distribution → ~25% greater 24-h muscle protein synthesis vs. evening-skewed (Mamerow, J Nutr)
`protein_spread`	Protein within zone but one meal contributes > 50% across the week	"Nearly all your protein comes from one meal. Spreading it out is less about the total and more about not depending on one dish."	as above
`nuts_handful`	No nuts or seeds in 7 days and magnesium or vitamin E below zone	"A small handful of nuts is about the densest easy win in nutrition."	Per 28 g/d: CHD RR 0.71, CVD 0.79, all-cause 0.78 (Aune 2016, BMC Medicine)
`fermented_add`	No fermented item (yoghurt, kefir, kimchi, sauerkraut, kombucha, miso, tempeh) in 7 days	"Yoghurt, kefir, sauerkraut — one fermented thing a day is a surprisingly well-studied habit."	17-week RCT: ↑ microbiota diversity, ↓ 19 inflammatory markers (Wastyk 2021, Cell)
`produce_volume`	Fibre, vitamin C and potassium all below zone on ≥3 of 7 days	"Risk keeps dropping up to about 800 g of fruit and veg a day — roughly five good handfuls. You're building toward it."	Dose–response to 800 g/d (Aune 2017, Int J Epidemiol)
`iron_vitc_pairing`	Iron below zone and top iron contributors are plant-dominant and vitamin C low in the same meals	"Iron from plants goes in better with something sour alongside — lemon over the lentils, peppers in the salad."	EFSA-authorised claim; meta-analysis ~+5.9 pp absorption (Proc Nutr Soc)
`calcium_add`	Calcium below zone on ≥4 of 7 days	"Dairy, fortified plant drinks, tofu, kale — and in Germany, some mineral waters carry more calcium than you'd expect."	DGE/USDA reference intakes; locale-aware DE variant
`omega3_add`	Omega-3 below zone and no oily fish, walnuts, flax or rapeseed oil in 7 days	"Oily fish, walnuts or a spoon of rapeseed oil — any one of them moves this a long way."	DGE/USDA guideline consensus
`sweet_drink_swap`	Sweetened drinks appear as ingredients on ≥3 days/week	"Trading one sweet drink for water is the single most-studied swap in nutrition. Just one — the rest can stay."	Water-for-SSB substitution → lower weight, lower obesity and T2D incidence (meta-analysis)
`salt_herbs`	Sodium above zone on ≥4 of 7 days	"Herbs, lemon and pepper do a lot of what salt does at the table. Worth keeping them within reach."	WHO/DGE sodium consensus, reframed additively per §5.2 rule 1
`post_meal_walk`	Activity context habitually low and the day's energy band is not above zone	"Ten minutes of walking after a meal does more than it sounds like — and it's the nicest ten minutes of the day."	Postprandial walking beats pre-meal or delayed activity (systematic review + meta-analysis, 2022)
`meal_regularity`	First-meal time varies by > 4 h across the week	"Your first meal has landed anywhere between 7 and 1 this week. Regular timing tends to make the rest of the day easier."	Consistency, not precision, is the strongest predictor across trials (July analysis §2)
`repeat_dish_anchor`	A `meal_memory` dish with ≥3 confirmations is a top contributor for a nutrient	"The lentil soup is quietly doing most of the work for your fibre. Nothing needs to change."	praise/anchor tip — the contribution pipeline at its simplest
`contribution_concentration`	One dish > 60% of a nutrient across the week	"Almost all your protein comes from one dish. Nothing wrong with that — just worth having a second option."	—
`light_days_care`	Energy band below zone on ≥3 consecutive days	"The last few days have been on the light side. Something a bit more substantial tomorrow would be kind to you."	Wellbeing-critical. Needs counsel sign-off and must never escalate — if it fires twice, it goes silent.
`first_steps`	< 3 logging days all-time	"Small start: just describe one meal tomorrow. That's the whole task."	Adherence literature — the smallest viable ask

5.4 Deliberate exclusions

Vitamin D. Almost every user will read below zone, and the honest advice — sunlight, possibly supplementation, possibly a blood test — is medical. Excluded from the library; show a neutral one-line explainer in All-details instead.

Any tip triggered by weight or weight stagnation. Absence and plateaus are ambiguous and reinstate weight as the success criterion (§4.5).

Any tip naming a mood→food correlation. Mood modulates tone and ambition only (§4.3).

Anything phrased as compensation. post_meal_walk in particular must never fire after a large meal or an above-zone energy band — that is the "work it off" mechanic wearing a health claim.

5.5 The two modifiers

Modifier	Effect
`low_energy_week` — tired/low mood tags on ≥2 of the last 3 logged days	Reduce the ambition of whichever tip fires and warm the register. Never mentioned in the copy.
`coaching_level` — from nutrition-knowledge level + logging history	Selects the copy variant: beginner tips name one concrete action; dialled-in tips can name the nutrient.

Why this matters commercially. Twenty deterministic tips cost nothing to run, cannot hallucinate, and can ship before the AI coaching epic. They also make the eventual AI layer cheaper and safer: the model receives a selected key plus the contributing dish and writes one sentence — instead of reasoning over a meal history and deciding what matters. Fewer tokens, less personal data in the prompt, bounded output.

06 · Further differentiation opportunities

Three sources: documented category complaints nobody has answered, the axes the comparison ecosystem scores on, and the gaps Apple's platform direction opens.

6.1 From unanswered category complaints

Opportunity	Why it is available	Cost
Eating out	The structural failure mode of every database app — restaurant food isn't in the database, and the standard advice is to google the restaurant's nutrition page or break the dish into ingredients yourself. Free text handles "the chicken tikka masala at the Indian place" natively. This already works; it simply isn't claimed.	Marketing only
Shared and home-cooked dishes	"We split a big pan of…" is unloggable in a database product and trivial in free text. Same situation: already true, never said.	Marketing only
Cuisine coverage	Food databases are structurally US/EU-centric; regional-dish coverage requires deliberate dataset work most apps have not done. An LLM handles Turkish, Levantine, Vietnamese or Nigerian home cooking far better than a crowdsourced database. In Germany this is a large, entirely unaddressed population, and no competitor markets to it.	Prompt evals + marketing
Cancellation as a feature	Noom carries >1,200 BBB cancellation complaints; MFP is in a class action over "free" marketing; the FTC re-opened click-to-cancel rulemaking in March 2026 and ~30 US states have their own auto-renewal laws. emealia's no-win-back-discount policy already exists — make one-tap cancellation visible.	Low
Alcohol without a moral event	Every tracker turns a beer into a budget breach. Nobody offers calm alcohol logging. Fits the non-negotiables exactly, no new mechanics.	Low
Export as a feature, not a checkbox	GDPR mandates it anyway. "Your data leaves whenever you want" counters lock-in and reinforces the privacy position.	Low
One price, shown	Direct counter to Yazio's documented price confusion in the primary market.	Marketing only

6.2 From the comparison ecosystem

Review and comparison lists score on: accuracy, speed, database size, micronutrient depth, integrations, price, coaching, offline, privacy. emealia loses by design on database size and integrations, and is unrankable on accuracy because it publishes ranges while everyone else publishes a false point estimate.

There is no column for "does this make you feel bad." So emealia will be scored on axes it deliberately loses, by reviewers who are mostly competing apps' content marketing.

The response is not to chase the lists — it is to create the axis. Publish a comparison scored on emealia's terms: red used in feedback · numbers forced on the main view · streaks · goal-weight countdowns · exercise-earns-calories mechanics · cancellation friction · data sold to third parties · features gated after purchase. Every one is objectively checkable, every mainstream competitor fails several, and a rubric is the classic way an outsider changes the terms of comparison — exactly what Yuka did to packaged food. It also gives the "clean creator" outreach a tool to hand their clients rather than a product to endorse.

6.3 From Apple's platform direction

Apple expands the funnel while commoditizing capture

Native nutrition logging will make casual food logging normal for a very large population — and simultaneously make "capture a meal" worth nothing. Value moves to interpretation. emealia is an interpretation product; it should stop competing on capture and say so explicitly.

Apple is iPhone-, English- and US-first

Android, desktop, German-language nuance and EU data residency are all outside its centre of gravity. A PWA covers every one by default. Android alone is roughly 60%+ of the German smartphone base.

Apple will ship rings, goals and numbers

Goal closure is the company's entire behavioral vocabulary. It will not ship "no numbers on the main view" — the single row in §3 that is safest from the platform.

Health+ raises the stakes on who holds the data

An AI reading your health data is about to be normalized. As the platform layer gets more intimate, "EU residency, one processor, never sold" gets more valuable, not less.

6.4 Ranked

Opportunity	Differentiation	Cost	Verdict
Claim eating-out + shared-dish + cuisine breadth	High	~zero	Do now — already true, never said
Guilt-design comparison rubric	High	Low	Do now — changes the axis instead of losing on theirs
Deterministic tip library (§5)	Medium-high	Low	Do now — ships before the AI coaching epic
Visible one-tap cancel + one price	Medium	Low	Do now
Market the number-hiding setting	Medium	~zero	Do now — already planned, never promoted
Weekly nutrient view	Medium-high	Medium	Strong candidate — a week is the honest unit for micronutrients, and weekly framing is inherently less guilt-inducing than daily. Nobody frames it that way.
Calm alcohol logging	Medium	Low	Candidate
Low-capacity / "rough day" logging mode	Medium	Medium	Candidate — fits non-negotiable #6
AI recipe generator	High	High	Plus, post-D7. The one place generative AI clearly earns its tokens; contributions are the ideal input; and a generated recipe is the only path to a logged meal with `confidence: high` and no estimation error. Requires hard allergen/exclusion constraint injection — a recipe is an instruction, not an estimate.
Garmin / Fitbit / Withings OAuth	Low-medium	Medium	Not now (§4.4)
Household / multi-eater	Low	High	No

07 · Where emealia stands

Genuine advantages

1 · The frame is uncontested. Ranges, no numbers, no red, no streaks, no verdict — still nobody in the mainstream. Supported by the science base in the July analysis (Frech 2022: precise ≈ round goals; Levinson 2017: 73.1% of affected MFP users saw the app as contributing to their eating disorder).

2 · Contributions + batch logging are two features nobody has, and both are consequences of the free-text architecture rather than bolt-ons.

3 · Price. €39.99/yr against Yazio's ~€83.90 and MFP's $79.99–99.99, everything included, no tiers.

4 · Privacy posture as product. EU residency, single processor, no data sale, no ad SDKs — against continued findings that health and fitness apps share dietary data with ad networks and brokers. In Germany this is a purchase argument, not a footnote.

5 · PWA independence. No App Store cut, no review risk, no gatekeeper, instant deploys.

Real exposures

1 · Discovery. No App Store means no store search, no charts, no default placement. 100% of acquisition is content, comparison SEO and creator outreach — and the "best app 2026" review ecosystem is largely owned by competing apps' content marketing. The biggest go-to-market risk, and no feature fixes it.

2 · Apple. Free native nutrition logging plus a paid AI coach, pre-installed. Everything generic is being commoditized.

3 · MFP + Cal AI. The "easy input" wedge is gone. Positioning must lead with the frame, not the ease.

4 · The German payment paradox is unchanged (~76% willing to use mHealth apps, ~27% willing to pay out of pocket; ~25% of German smartphone users have a weight/nutrition app installed). Still the central bet.

5 · Coaching maturity gap — see §4.6. 6 · No wearable/health-platform sync, structurally limited (§4.4).

7 · Self-inflicted: the public landing page still advertises a 7-day trial while strategy and the shipped UsageMeter implement 30 AI credits. With a competitor currently facing a class action over inaccurate "free" marketing, this is a real risk, not a copy nit. Fix before any acquisition spend.

08 · Recommendations

#	Recommendation
1	Re-lead the positioning on the frame, not the ease. "Meal tracking made easy" is now a claim MFP can match. "Eat by feel, not by math" cannot be matched by anyone whose product is a number. Rewrite the landing hero hierarchy so guilt-free comes first and easy comes second.
2	Make contributions a headline feature, and gate it on the copy. Add it to the G1 comprehension test — if testers read the list as blame, do not ship it as written. Put ordering, labelling and "still estimating" rules into acceptance criteria, not just design.
3	Market multi-meal logging explicitly. "Log your whole day in one sentence" is unique in the field today and directly answers the category's best-documented complaint. Landing page and demo.
4	Build the contribution → suggestion pipeline as the primary coaching path. Cheaper, more specific, less personal data in the prompt, structurally incapable of shaming.
5	Ship the Yazio and MyFitnessPal comparison pages first. Yazio at ~2.1× the price with a worse input method in the primary market; MFP with a live class action over paywalled "free" features. Both write themselves and both are honest.
6	Write the "what we will not build" list down as strategy. No GLP-1 features, no goal-weight countdown, no earned-calories from activity, no mood↔food causality, no adaptive TDEE from weight trend, no barcode scanner. Each will be proposed by someone; each contradicts a non-negotiable; and the list itself is a differentiator worth publishing in a lighter form.
7	Add an EU AI Act Art. 50 line item next to the GDPR epic, and use it in DE marketing alongside DSGVO. emealia already complies in substance; it needs the disclosure placement and the sentence.
8	Fix the trial mismatch on the public site before spending anything on acquisition.
9	Treat fitness-app sync as a scoped, honest decision, not an open wish: Apple Health is unavailable to a PWA; Garmin/Fitbit/Withings/Strava are feasible over OAuth; burned calories never credit the energy band.
10	Instrument contributions through `UsageMeter` from day 1 as a premium candidate, per the existing rule — even though it costs nothing to run. If it turns out to be the feature people open the app for, that is worth knowing before Plus scope is decided.
11	Ship the deterministic tip library (§5) before the AI coaching epic. Twenty evidence-backed rules cost nothing, cannot hallucinate, are auditable under AI Act Art. 50, and make the eventual AI layer cheaper and safer by reducing it to a phrasing job.
12	Claim what is already true (§6.1): eating out, shared dishes, and cuisine breadth beyond the US/EU-centric database canon. Zero build cost, high differentiation, and in Germany the cuisine point reaches a large population no competitor is addressing.
13	Publish a guilt-design comparison rubric (§6.2). The review ecosystem has no column for "does this make you feel bad", so emealia is scored only on axes it loses by design. Create the axis instead of chasing theirs — and hand it to the creator outreach as a tool rather than an endorsement.
14	Scope a weekly nutrient view. A week is the honest unit for micronutrients, weekly framing is structurally less guilt-inducing than daily, and nobody in the field frames it that way.

09 · Watchlist

Signal	Threshold	Response
Apple Health+ ships a range-based or non-judgmental mode	any	Re-examine the frame's defensibility immediately — the one thing that would erase the moat
A mainstream tracker ships retrospective, non-numeric contribution	any	Novelty window on §4.1 closes; shift emphasis to batch logging + frame
MFP paywall class action	resolution	Either a marketing gift or a reduced argument; watch either way
Trial → paid	<5%	Per the July benchmark: rework onboarding and value prop before scaling marketing
Reviews / support mentioning obsession or anxiety	recurring	Reduce number visibility further; re-audit the contribution copy first — likeliest source
German nutrition-app price movement	Yazio below ~€50/yr	Primary-market price argument weakens; lean harder on privacy + frame

10 · Caveats

The 2026 "best app" and "accuracy benchmark" ecosystem is advertising. Nutrola, PlateLens, Welling, NutriScan, Amy Food Journal, Fitia, Hoot and similar all sell competing apps. Accuracy rankings from those sources are marketing claims and are treated as such here. Only peer-reviewed work and first-party or news-outlet reporting are treated as evidence.
Apple Health+ is pre-release reporting, repeatedly delayed in the rumor cycle. Directional only.
Market-size estimates vary by a factor of 2–7 depending on definition ($2.8 bn to $16.9 bn for 2026 across firms). Useful for direction, useless as a number.
Pricing changes constantly and regionally (Yazio's German pricing in particular varies by entry point); figures are August 2026 snapshots.
MacroFactor's $89.99/yr "bundle" pricing appears in secondary sources without a consistent description of what is bundled; treat as indicative.
The 2026-07-07 analysis remains authoritative for the science base, the German willingness-to-pay data and the token/COGS model. This document does not restate them.

Sources of truth. Full markdown version with all inline source links: emealia-mgmt/market-analysis/2026-08-competitor-analysis.md. Predecessor: compass_artifact_wf-38ad3532-…md (2026-07-07). On any conflict about product concept, emealia-mgmt/ wins.