<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[Mealumo Dev Notes]]></title><description><![CDATA[Mealumo Dev Notes]]></description><link>https://mealumo.hashnode.dev</link><generator>RSS for Node</generator><lastBuildDate>Tue, 03 Nov 2026 06:21:45 GMT</lastBuildDate><atom:link href="https://mealumo.hashnode.dev/rss.xml" rel="self" type="application/rss+xml"/><language><![CDATA[en]]></language><ttl>60</ttl><item><title><![CDATA[Cómo construimos un registro de comidas por foto para iOS (y por qué la IA no ve el aceite)]]></title><description><![CDATA[De la foto del plato a una entrada editable: por qué la pantalla de revisión importa más que el modelo]]></description><link>https://mealumo.hashnode.dev/c-mo-construimos-un-registro-de-comidas-por-foto-para-ios-y-por-qu-la-ia-no-ve-el-aceite</link><guid isPermaLink="true">https://mealumo.hashnode.dev/c-mo-construimos-un-registro-de-comidas-por-foto-para-ios-y-por-qu-la-ia-no-ve-el-aceite</guid><category><![CDATA[iOS]]></category><category><![CDATA[AI]]></category><category><![CDATA[spanish]]></category><category><![CDATA[Mobile Development]]></category><dc:creator><![CDATA[Vlad]]></dc:creator><pubDate>Sun, 25 Oct 2026 09:03:57 GMT</pubDate><content:encoded><![CDATA[<p><em>Publicado originalmente en dev.to: <a href="https://dev.to/mealumo/como-construimos-un-registro-de-comidas-por-foto-para-ios-y-por-que-la-ia-no-ve-el-aceite-8jk">Cómo construimos un registro de comidas por foto para iOS (y por qué la IA no ve el aceite)</a>. Lo republico aquí con un enlace al original.</em></p>
<blockquote>
<p><strong>Aviso:</strong> soy el desarrollador de Mealumo, la app de iPhone de la que habla este post. No es una reseña, sino un repaso de las decisiones de diseño detrás de ella. <strong>La app está solo en inglés</strong> y requiere <strong>iOS 26 o posterior</strong>. Este texto se ha redactado con ayuda de IA y lo he revisado antes de publicarlo.</p>
</blockquote>
<h2>El punto de partida</h2>
<p>Cuando alguien abandona una app para contar calorías, rara vez es por la nutrición. Es por el tecleo. Escribes "ensalada de pollo", te salen decenas de entradas casi iguales, adivinas la cantidad y repites el proceso con el aliño, el pan y la bebida. Hacer eso en cada comida cansa a cualquiera.</p>
<p>La idea de Mealumo es sencilla: que la entrada empiece con una foto. Fotografías el plato (o eliges una imagen de la galería), la app propone ingredientes, cantidades, calorías, proteínas, hidratos y grasas, y tú arreglas lo que no cuadra antes de guardarlo.</p>
<p>Esa última parte, la de arreglar, resultó ser el producto de verdad.</p>
<h2>El flujo, paso a paso</h2>
<ol>
<li><strong>Captura.</strong> Con la cámara o desde la galería. Poder importar desde la galería permite registrar por la noche lo que fotografiaste a mediodía.</li>
<li><strong>Análisis.</strong> La imagen se envía a un servicio de IA externo, que devuelve su estimación: qué alimentos cree que hay, cuánto de cada uno y sus macros. Este paso necesita conexión a internet.</li>
<li><strong>Revisión.</strong> El resultado aparece como una lista editable, no como una cifra cerrada. Puedes cambiar cantidades, editar ingredientes y añadir lo que falte mediante el buscador de alimentos.</li>
<li><strong>Guardado.</strong> La entrada pasa al diario del día, que suma calorías y macros y las compara con tus objetivos diarios.</li>
</ol>
<p>Alrededor de ese núcleo están las piezas menos vistosas: registro de agua, registro de peso con su evolución, ideas de recetas y objetivos diarios estimados a partir de tus metas, medidas y nivel de actividad.</p>
<h2>Por qué una sola cifra es mala idea</h2>
<p>La versión tentadora de esta función es la de la demo: entra una foto y sale "640 kcal". Queda espectacular, pero falla de una forma que el usuario no puede arreglar.</p>
<p>Si la app dice 640 y tú sabes que la ración era mayor, ¿qué tocas? Si sobrescribes el total, las macros dejan de sumar, y sigues sin saber qué parte del plato estaba mal.</p>
<p>Desglosar la estimación en <strong>alimentos con su cantidad</strong> lo resuelve. Los fallos en una foto de comida suelen ser localizados: se calcula mal el arroz, o la salsa resulta ser de nata y no de tomate. Con una lista, corriges esa línea y los totales se recalculan solos. El modelo no tiene que acertar siempre. Le basta con ser mejor punto de partida que un buscador vacío.</p>
<h2>Lo que una foto no puede contar</h2>
<p>El límite no es solo el modelo. Hay información que, sencillamente, no está en los píxeles:</p>
<ul>
<li><strong>La grasa de cocinar.</strong> Una cucharada de aceite de oliva ronda las 120 kcal y desaparece en cuanto se integra en el plato.</li>
<li><strong>La profundidad.</strong> Desde arriba se ve la superficie, no el volumen. Un cuenco hondo y un plato llano pueden parecer iguales.</li>
<li><strong>Lo que va por dentro.</strong> Burritos, empanadas, lasañas: todo lo relleno o mezclado.</li>
<li><strong>La receta.</strong> "Pasta con salsa" puede ser algo ligero o algo muy contundente.</li>
</ul>
<p>Esto no se arregla con más ingeniería. Lo que sí se puede hacer es decirlo con claridad y que corregirlo cueste poco. Por eso la ficha de la App Store avisa de que las estimaciones pueden ser inexactas y conviene revisarlas, y por eso la revisión forma parte del flujo en lugar de ser un extra opcional.</p>
<h2>Rapidez frente a confianza</h2>
<p>Hay una tensión real: registrar solo merece la pena si es más rápido que teclear, y cada confirmación añade fricción. Estos son los principios con los que diseñamos:</p>
<ul>
<li><strong>Mostrar la estimación al momento y que sea editable.</strong> Lo primero que ves tras el análisis tiene que permitirte actuar.</li>
<li><strong>Tratar la cantidad como lo principal a corregir</strong>, porque es lo que peor se deduce de una imagen.</li>
<li><strong>Tener la entrada manual a mano</strong> para las comidas que no se dejan fotografiar bien, como un batido en una botella opaca.</li>
<li><strong>No aparentar más precisión de la que hay.</strong> Una estimación tiene que parecer una estimación.</li>
</ul>
<h2>Privacidad, en la ficha y no en la letra pequeña</h2>
<p>Las fotos de comida son personales, y mandarlas a un servicio de terceros es una decisión que el usuario debe conocer. La ficha de Mealumo en la App Store lo dice sin rodeos: las fotos de las comidas y los datos de perfil necesarios para calcular los objetivos se envían a un servicio de IA externo para su procesamiento. Es poca cosa, pero en una app cercana a la salud ese tipo de detalle pesa en la confianza.</p>
<h2>No es una app médica</h2>
<p>Mealumo es un diario de comidas de bienestar general. Los objetivos son estimaciones, no diagnostica ni trata nada, y así lo indica la ficha. No promete resultados de salud, y creo que ninguna app de este tipo debería hacerlo.</p>
<h2>Si trabajas con modelos de visión</h2>
<ul>
<li><strong>Devuelve estructura, no un veredicto.</strong> Una salida desglosada permite corregir una parte sin tirar el resto.</li>
<li><strong>Ten claro qué no puede contener tu entrada.</strong> En comida son la grasa, la profundidad y lo que queda oculto; cada dominio tiene lo suyo.</li>
<li><strong>Haz visible la incertidumbre en la interfaz.</strong> Un resultado editable comunica "esto es una estimación" mejor que cualquier aviso legal.</li>
<li><strong>Explica adónde van los datos.</strong> La gente se fija.</li>
</ul>
<h2>Pruébala y dime dónde falla</h2>
<p>Mealumo es de descarga gratuita para iPhone (iOS 26+, app en inglés), con una suscripción PRO opcional para seguir escaneando y acceder a todas las recetas:
<a href="https://apps.apple.com/app/id6811904664">https://apps.apple.com/app/id6811904664</a></p>
<p>Cualquier plan PRO empieza con 3 días de prueba gratis. Si quieres probarlo más tiempo, deja un comentario o escríbeme por mensaje directo y te paso un código promocional PRO gratis. Y me interesa especialmente lo contrario: fotos en las que la estimación se vaya claramente.</p>
<p>¿Has trabajado con modelos de visión sobre datos reales y desordenados? ¿Cómo gestionaste los casos en los que la respuesta simplemente no está en la imagen?</p>
]]></content:encoded></item><item><title><![CDATA[Logging meals when half of them aren't yours to measure]]></title><description><![CDATA[Disclosure: I'm the developer of Mealumo, an iPhone app that estimates calories from a meal photo. It comes up once, at the end. This essay was drafted…]]></description><link>https://mealumo.hashnode.dev/logging-meals-when-half-of-them-aren-t-yours-to-measure</link><guid isPermaLink="true">https://mealumo.hashnode.dev/logging-meals-when-half-of-them-aren-t-yours-to-measure</guid><category><![CDATA[Nutrition]]></category><category><![CDATA[iOS]]></category><category><![CDATA[apps]]></category><category><![CDATA[Productivity]]></category><dc:creator><![CDATA[Vlad]]></dc:creator><pubDate>Sat, 24 Oct 2026 08:03:11 GMT</pubDate><content:encoded><![CDATA[<p><em>Disclosure: I'm the developer of Mealumo, an iPhone app that estimates calories from a meal photo. It comes up once, at the end. This essay was drafted with AI assistance and reviewed and edited by me.</em></p>
<p><em>Originally published on my Bear Blog: <a href="https://mealumo.bearblog.dev/logging-meals-when-half-of-them-arent-yours-to-measure/">https://mealumo.bearblog.dev/logging-meals-when-half-of-them-arent-yours-to-measure/</a></em></p>
<p>Most advice about tracking food quietly assumes you cook every meal from a recipe, with a scale on the counter. Real weeks don't look like that. There's the Tuesday shawarma plate, the potluck at a friend's place, a double-double on the drive in, and the chili your partner made without measuring anything. A good chunk of what you eat comes from kitchens where nobody wrote anything down.</p>
<p>That's not a reason to give up on logging. It just means the goal has to change.</p>
<h2>Stop aiming for exact</h2>
<p>When you didn't cook it, you can't know the butter, the oil, or the sugar in the sauce. Pretending otherwise leads to one of two outcomes: you spend ten minutes hunting for a database entry that matches "Mom's lasagna," or you skip the meal entirely. The second is worse. A log full of gaps around restaurant dinners and family meals tells you less than a rough one that includes them.</p>
<p>So aim for "close enough, recorded." A reasonable guess written down beats a perfect number you never entered.</p>
<h2>Eating out: log the plate, then the extras</h2>
<p>Restaurant portions vary, but the plate itself is usually readable: a protein, a starch, some vegetables, a sauce. Log the main components first. Then think briefly about what you can't see. Was it fried? Was there a buttery finish on the vegetables? Was the dressing already tossed in? Add a rough amount for those. It takes seconds and it's often the biggest single correction.</p>
<p>Drinks deserve their own line. A latte, a pint, or a pop can carry as much as a side dish, and they're the easiest thing to forget.</p>
<h2>Home cooking from someone else's hands</h2>
<p>Shared home meals are harder in a different way. You often know the ingredients but not the amounts. A few habits help:</p>
<ul>
<li><strong>Ask one question, not ten.</strong> "Was there cream in this?" is a normal dinner-table question. A full ingredient interview isn't.</li>
<li><strong>Use your own cooking as a reference.</strong> If you've made a similar stew, use your version as the starting point and adjust.</li>
<li><strong>Log what you served yourself</strong>, not the whole pot. Your portion is the part you actually saw.</li>
</ul>
<h2>Mixed days are normal days</h2>
<p>A typical day might be oatmeal you made, a sandwich from the café near work, and a dinner someone else cooked. Treat each meal with the precision it allows: careful where you can be, rough where you can't. Over a week, the pattern still shows up, and the pattern is the useful part. Single meals will be off in both directions; that's fine.</p>
<p>If tracking ever starts to feel stressful rather than informative, it's okay to step back. And for specific medical needs, a doctor or registered dietitian is the right guide, not an app.</p>
<h2>Where a photo helps</h2>
<p>This is the case I had in mind when building Mealumo. On a night out or at a friend's table, taking a quick photo is less awkward than typing in five ingredients, and you can log it later from your camera roll. The app estimates the ingredients, portions, calories, protein, carbs and fat, and then shows you a review screen where you fix what it got wrong: add the oil, shrink the portion, swap the ingredient it misread. It also has a food journal, food search, water and weight logs, and recipe ideas.</p>
<p>It's an estimate, not a measurement, and it's built around you correcting it.</p>
<p>Mealumo is free to download on iPhone (iOS 26+), with an optional PRO subscription; every PRO plan starts with a 3-day free trial: <a href="https://apps.apple.com/app/id6811904664">https://apps.apple.com/app/id6811904664</a></p>
<p>If you'd like to try PRO, get in touch and I'll send you a free PRO promo code. And if it badly misjudges your poutine, I genuinely want to hear about it.</p>
]]></content:encoded></item><item><title><![CDATA[Shipping an iOS 26-only app in 2026: how do you decide where the cutoff goes?]]></title><description><![CDATA[Originally published on dev.to: Shipping an iOS 26-only app in 2026: how do you decide where the cutoff goes?. Republished here with a canonical link to…]]></description><link>https://mealumo.hashnode.dev/shipping-an-ios-26-only-app-in-2026-how-do-you-decide-where-the-cutoff-goes</link><guid isPermaLink="true">https://mealumo.hashnode.dev/shipping-an-ios-26-only-app-in-2026-how-do-you-decide-where-the-cutoff-goes</guid><category><![CDATA[Mobile Development]]></category><category><![CDATA[discussion]]></category><category><![CDATA[iOS]]></category><category><![CDATA[Swift]]></category><dc:creator><![CDATA[Vlad]]></dc:creator><pubDate>Fri, 23 Oct 2026 08:09:57 GMT</pubDate><content:encoded><![CDATA[<p><em>Originally published on dev.to: <a href="https://dev.to/mealumo/shipping-an-ios-26-only-app-in-2026-how-do-you-decide-where-the-cutoff-goes-4862">Shipping an iOS 26-only app in 2026: how do you decide where the cutoff goes?</a>. Republished here with a canonical link to the original.</em></p>
<p><em>Disclosure: I'm the developer of Mealumo, an iPhone calorie-tracking app. I'm using it as the example here, but this is a question for the community, not a launch post.</em></p>
<p>Mealumo requires iOS 26.0 or later. That's the minimum on its App Store page, and it means anyone who hasn't updated to this year's iOS can't install it.</p>
<p>I'm not going to argue that this was the right or wrong call. I'd rather hear how other people make this decision, because it seems to be one of those choices every new app makes and few people talk about openly.</p>
<h2>What the listing actually says</h2>
<p>For context, here's the compatibility section on the App Store page:</p>
<ul>
<li><strong>iPhone:</strong> requires iOS 26.0 or later</li>
<li><strong>Mac:</strong> requires macOS 26.0 or later and a Mac with an Apple M1 chip or later</li>
<li><strong>Apple Vision:</strong> requires visionOS 26.0 or later</li>
</ul>
<p>So the cutoff carries over to other platforms too, which raises its own questions.</p>
<h2>Questions I'd love your take on</h2>
<p><strong>1. How do you pick a minimum version for a brand-new app?</strong>
Do you start from the latest OS and only go back if you need to? Or start with N-1 or N-2 and only drop versions when an API forces you? Is there a rule of thumb you trust?</p>
<p><strong>2. What data do you actually look at?</strong>
Apple publishes adoption figures from time to time, and analytics vendors publish their own. Which sources do you trust, and do you weight them differently for a niche audience? For a consumer health and fitness app, for example, would you expect users to update faster or slower than average?</p>
<p><strong>3. Which APIs have pushed you to require the newest iOS?</strong>
I'm curious whether it's usually one specific framework, a SwiftUI behavior that only works properly on the latest release, or simply the cost of maintaining <code>if #available</code> branches.</p>
<p><strong>4. Does the cutoff show up in your numbers?</strong>
If you've shipped a latest-only app, did you see a measurable difference in product page conversion, or in reviews from people who couldn't install it? Or was it hard to notice at all?</p>
<p><strong>5. Would you lower the minimum after launch?</strong>
Going back to support older versions later means testing on more OS versions and possibly rewriting features. Has anyone done this, and was it worth it?</p>
<p><strong>6. How does this play out on Mac and visionOS?</strong>
When an iPhone app becomes available on Apple silicon Macs or Vision Pro, do you think about those minimums at all, or do they just follow whatever you chose for iOS?</p>
<h2>Why I'm asking</h2>
<p>The choice about how far back to support older iOS versions tends to get made early and then stays in place for a long time. I'd rather understand how experienced iOS devs think about it than assume there's one obvious answer.</p>
<p>Please share your reasoning, your data sources, or the times you got it wrong. Disagreement is very welcome.</p>
<hr />
<p>If you're on iOS 26 and want to see the app in question, it's here: <a href="https://apps.apple.com/app/id6811904664">https://apps.apple.com/app/id6811904664</a>. It's free to download, any PRO plan starts with a 3-day free trial, and if you want to try PRO longer, comment or DM me for a free PRO promo code.</p>
<p><em>This post was drafted with AI assistance and reviewed before publishing.</em></p>
]]></content:encoded></item><item><title><![CDATA[Most people don't quit calorie counting because of the math]]></title><description><![CDATA[The hard part is logging. Here's what photo-based logging fixes, what it gets wrong, and how to get more honest numbers out of it.]]></description><link>https://mealumo.hashnode.dev/most-people-don-t-quit-calorie-counting-because-of-the-math</link><guid isPermaLink="true">https://mealumo.hashnode.dev/most-people-don-t-quit-calorie-counting-because-of-the-math</guid><category><![CDATA[Nutrition]]></category><category><![CDATA[Health,]]></category><category><![CDATA[habits]]></category><category><![CDATA[AI]]></category><category><![CDATA[iOS]]></category><dc:creator><![CDATA[Vlad]]></dc:creator><pubDate>Tue, 20 Oct 2026 08:09:35 GMT</pubDate><enclosure url="https://cdn.hashnode.com/uploads/covers/6ad5e3dadd1639045b61bc68/14d63111-e411-4705-87a3-d18bb988a308.png" length="0" type="image/jpeg"/><content:encoded><![CDATA[<p><em>Originally published on</em> <a href="https://buildingmealumo.substack.com/p/most-people-dont-quit-calorie-counting"><em>Building Mealumo</em></a><em>, my Substack. Republished here with a canonical link to the original.</em></p>
<p><strong>Disclosure:</strong> I'm the developer of Mealumo, an AI calorie-tracking app, so I have a stake in this topic. I've tried to make this useful whether or not you ever download it. The app comes up once, near the end.</p>
<p><em>This piece was drafted with AI assistance and reviewed and edited by me before publishing.</em></p>
<hr />
<p>Ask people why they stopped counting calories and they rarely say the numbers were confusing. Usually it just got annoying.</p>
<p>The arithmetic is easy. The friction is in the logging: finding the right entry among dozens of near-duplicates, guessing whether that was 150 or 250 grams of rice, remembering the dressing, the milk in your coffee, the handful of nuts at 4 pm. Each one takes a minute. Multiply that by every meal and snack, and tracking becomes a chore that's easy to quietly drop.</p>
<h2>The useful goal is "roughly right, most days"</h2>
<p>If you track to become more aware of what you eat, an approximate log you keep for months tells you more than a perfect one you keep for three days. Gaps tend to hurt more than small errors, because the meals that go unlogged are often the ones that matter: the restaurant dinner, the afternoon snack, the drink.</p>
<p>That changes the question. Instead of "what's the most accurate method?", ask "what's the most accurate method I'll actually keep using?"</p>
<p>For many people, that's the one with the fewest taps.</p>
<h2>What a photo is good at</h2>
<p>Taking a picture of your plate is about as low-effort as logging gets. Your phone is probably already in your hand. Photo-based trackers use an AI model to look at the image and estimate what's on the plate, how much of each item, and the calories and macros (protein, carbs, fat).</p>
<p>It works best for:</p>
<ul>
<li><p><strong>Plates with distinct items.</strong> Chicken, rice and broccoli side by side is the easy case.</p>
</li>
<li><p><strong>Restaurant meals</strong>, where you don't have a recipe and would be guessing anyway.</p>
</li>
<li><p><strong>Logging later.</strong> Snap it now and log it from your camera roll tonight.</p>
</li>
<li><p><strong>Small things.</strong> A photo takes a couple of seconds, so the snack actually gets recorded.</p>
</li>
</ul>
<h2>What a photo gets wrong (and always will)</h2>
<p>Apps in this space, mine included, should be upfront about this. Some calories just aren't visible:</p>
<ul>
<li><p><strong>Cooking oil and butter.</strong> A tablespoon of oil is roughly 120 kcal and disappears into the food.</p>
</li>
<li><p><strong>Sauces and dressings</strong>, especially creamy ones, and how much was used.</p>
</li>
<li><p><strong>Depth.</strong> A photo from above shows how wide the food is, not how deep the bowl is.</p>
</li>
<li><p><strong>Mixed and wrapped food</strong>: burritos, casseroles, sandwiches, smoothies.</p>
</li>
<li><p><strong>Drinks.</strong> A sweet latte and a black coffee can look much the same in a cup.</p>
</li>
</ul>
<p>None of this is a reason to skip photo logging. It means treating the estimate as a first draft, not a final answer.</p>
<h2>Five habits that make photo estimates more honest</h2>
<ol>
<li><p><strong>Shoot at an angle, around 45 degrees</strong>, rather than straight down, so the depth is easier to judge.</p>
</li>
<li><p><strong>Photograph before you mix.</strong> Take the salad before the dressing goes on and the bowl before you stir it.</p>
</li>
<li><p><strong>Add the invisible stuff by hand.</strong> If you cooked with oil, add it. It takes a few seconds and is often the biggest correction you'll make.</p>
</li>
<li><p><strong>Check the portion first.</strong> It's what a photo shows least reliably, so look there before anything else.</p>
</li>
<li><p><strong>Look at the week, not the meal.</strong> Single meals will be off in both directions. The weekly pattern is what tells you something.</p>
</li>
</ol>
<h2>What a tracker can't do for you</h2>
<p>A tracker gives you information. It doesn't make decisions for you, and it isn't medical advice. Calorie and macro targets from any app are estimates based on what you tell it. If you have a medical condition, a history of disordered eating, or specific health goals, a doctor or registered dietitian is a better guide than any app. And if counting starts to make you anxious rather than informed, it's okay to stop.</p>
<h2>Where Mealumo fits</h2>
<p>That's the thinking behind Mealumo, the iPhone app I develop. You take a photo of a meal or pick one from your library. It estimates the ingredients, portions, calories, protein, carbs and fat, and you review and adjust everything before saving. It also logs water and weight and has recipe ideas. It doesn't claim to be exact. The review step is there because no photo estimate can be.</p>
<p>It's free to download on iPhone (iOS 26+), with an optional PRO subscription: <a href="https://apps.apple.com/app/id6811904664">https://apps.apple.com/app/id6811904664</a></p>
<p>It's new and has no reviews yet. If you try it and the estimate for your dinner is way off, I'd really like to know. And if you'd like to try PRO, comment or DM me for a free PRO promo code.</p>
]]></content:encoded></item><item><title><![CDATA[Estimer les calories d'un repas à partir d'une photo : 4 méthodes comparées]]></title><description><![CDATA[De la paume de la main à l'application dédiée : ce que chaque méthode fait bien, ses limites, et comment en choisir une que vous garderez]]></description><link>https://mealumo.hashnode.dev/estimer-les-calories-d-un-repas-partir-d-une-photo-4-m-thodes-compar-es</link><guid isPermaLink="true">https://mealumo.hashnode.dev/estimer-les-calories-d-un-repas-partir-d-une-photo-4-m-thodes-compar-es</guid><category><![CDATA[French]]></category><category><![CDATA[Nutrition]]></category><category><![CDATA[Health,]]></category><category><![CDATA[AI]]></category><category><![CDATA[apps]]></category><dc:creator><![CDATA[Vlad]]></dc:creator><pubDate>Mon, 19 Oct 2026 10:30:34 GMT</pubDate><content:encoded><![CDATA[<p><em>Transparence : je suis le développeur de Mealumo, l'une des applications citées plus bas. J'ai voulu que ce guide vous soit utile quel que soit votre choix, y compris les options gratuites qui se passent de toute application. À noter : Mealumo est uniquement en anglais et nécessite iOS 26 ou une version ultérieure.</em></p>
<p><em>Cet article a été rédigé avec l'aide de l'IA et relu avant publication.</em></p>
<p>L'assiette est devant vous, vous aimeriez avoir une idée des calories, mais pas de balance de cuisine à portée de main. Peut-être êtes-vous au restaurant, ou vous n'avez tout simplement pas envie de peser les pâtes. Voici quatre façons concrètes d'évaluer ce qu'il y a dans l'assiette, du « sans aucun outil » à l'application spécialisée, avec leurs points forts et leurs limites.</p>
<p><strong>Avant de commencer.</strong> Les quatre méthodes donnent une estimation, pas une mesure. Ni un humain ni un logiciel ne voient ce que contient un gratin ou ce qui se cache sous la sauce. Pour se faire une idée générale, cela suffit. Si vous avez besoin de chiffres exacts pour des raisons médicales, pesez vos aliments et faites-vous accompagner par un ou une diététicien(ne) ou par votre médecin.</p>
<h2>Méthode 1 : les portions avec la main (gratuit, sans téléphone)</h2>
<p>Votre main est un instrument de mesure que vous avez toujours sur vous, et elle est grosso modo proportionnée à votre corps. Une règle empirique très répandue :</p>
<ul>
<li><strong>Paume</strong> (taille et épaisseur) : une portion de protéines, par exemple viande, poisson, tofu ou œufs</li>
<li><strong>Poing</strong> : une portion de légumes</li>
<li><strong>Main en coupe</strong> : une portion de glucides, par exemple riz, pâtes, céréales ou fruits</li>
<li><strong>Pouce</strong> : une portion de matières grasses, par exemple huile, beurre, beurre de cacahuète ou fromage</li>
</ul>
<p>Comptez les portions dans l'assiette, multipliez chacune par une valeur calorique typique de sa catégorie, et vous obtenez un ordre de grandeur.</p>
<p><strong>Utile pour :</strong> développer son intuition, voyager, ou les situations où sortir son téléphone serait malvenu. <strong>À garder en tête :</strong> c'est approximatif, cela demande un peu d'entraînement, et les « paumes » ont tendance à grandir sans qu'on s'en aperçoive.</p>
<h2>Méthode 2 : un objet de référence et une table de composition (gratuit)</h2>
<p>Photographiez le plat avec un objet de taille connue à côté : une fourchette, une canette ou une carte bancaire. Servez-vous-en pour évaluer la quantité de chaque élément, puis cherchez chaque aliment dans une table de composition nutritionnelle. En France, la table Ciqual de l'Anses est gratuite. L'USDA FoodData Central, aux États-Unis, en est une autre.</p>
<p><strong>Utile pour :</strong> des vérifications ponctuelles et soigneuses, avec des données de référence fiables. <strong>À garder en tête :</strong> il faut plusieurs minutes par repas, ce qui rend la méthode difficile à tenir au quotidien.</p>
<h2>Méthode 3 : un assistant IA généraliste (souvent gratuit)</h2>
<p>Plusieurs assistants IA capables d'analyser des images peuvent décrire la photo d'un plat et en estimer les apports si vous le leur demandez. Une consigne qui fonctionne bien : <em>« Liste chaque aliment visible, estime la quantité de chacun, indique calories, protéines, glucides et lipides par élément, et précise tes hypothèses. »</em></p>
<p><strong>Utile pour :</strong> une question ponctuelle, et pour comprendre le raisonnement derrière un chiffre. <strong>À garder en tête :</strong> pas de journal, pas d'objectifs, pas d'historique ; la réponse peut changer si vous reposez la question, et c'est à vous de noter les chiffres ailleurs.</p>
<h2>Méthode 4 : une application dédiée au suivi par photo</h2>
<p>Ces applications associent l'estimation à partir de la photo à un journal alimentaire, des objectifs quotidiens et un historique. Quelques exemples, d'après leurs descriptions sur l'App Store en octobre 2026 (les fonctionnalités et les formules évoluent, vérifiez donc la fiche actuelle avant de choisir) :</p>
<ul>
<li><strong>MyFitnessPal :</strong> un compteur de calories de longue date. Le scan de repas (meal scan) figure parmi les fonctionnalités Premium, avec le scan de codes-barres et la saisie vocale.</li>
<li><strong>Lose It! :</strong> une autre application installée depuis longtemps. L'enregistrement des repas par photo fait partie des fonctionnalités de sa formule Premium.</li>
<li><strong>Cal AI :</strong> conçue autour de la photo du repas. Sa fiche précise que les résultats de l'analyse des aliments nécessitent un abonnement.</li>
<li><strong>SnapCalorie :</strong> saisie par photo ou par note vocale ; l'application estime les aliments et les portions.</li>
<li><strong>Mealumo</strong> (celle que je développe) : iPhone uniquement, iOS 26 ou ultérieur, <strong>application en anglais</strong>. Vous prenez une photo ou en choisissez une dans la photothèque. L'app estime ingrédients, portions, calories, protéines, glucides et lipides, et vous vérifiez et modifiez le résultat avant d'enregistrer. On y trouve aussi un journal alimentaire, une recherche d'aliments, le suivi de l'eau et du poids, et des idées de recettes. Le téléchargement est gratuit, avec un abonnement PRO facultatif pour continuer à scanner. L'analyse des photos passe par un service d'IA externe et nécessite une connexion internet.</li>
</ul>
<p><strong>Utile pour :</strong> en faire une habitude, puisqu'une photo suffit et que l'historique s'étoffe jusqu'à faire apparaître des tendances. <strong>À garder en tête :</strong> les formules payantes, une précision qui varie selon les plats, et tout ce que l'appareil photo ne capte pas reste à ajouter vous-même.</p>
<p><strong>Comment choisir :</strong> essayez-en quelques-unes pendant plusieurs jours avec vos repas habituels. Gardez celle dont les résultats vous paraissent les plus simples à <em>vérifier et corriger</em>, plutôt que celle au marketing le plus tapageur.</p>
<h2>Pour une meilleure estimation, quelle que soit la méthode</h2>
<ul>
<li><strong>Laissez de la marge autour de l'assiette.</strong> Une photo trop serrée coupe l'accompagnement, la sauce et le verre posé à côté.</li>
<li><strong>Privilégiez une bonne lumière plutôt qu'un bel angle.</strong> Les ombres et la pénombre de nombreux restaurants rendent les aliments difficiles à distinguer.</li>
<li><strong>Ajoutez ce que vous savez.</strong> Si vous avez cuisiné, vous savez si le lait était entier ou écrémé et combien de beurre il y avait. Précisez-le.</li>
<li><strong>D'abord le plat, ensuite les extras.</strong> Notez le plat principal, puis revenez sur ce qui a été ajouté : vinaigrette, sauce, sucre dans le café.</li>
<li><strong>Restez fidèle à une méthode.</strong> En changer tous les deux jours rend vos chiffres difficiles à comparer.</li>
</ul>
<h2>En un coup d'œil</h2>
<table>
<thead>
<tr>
<th>Méthode</th>
<th>Coût</th>
<th>Effort par repas</th>
<th>Ce que vous obtenez</th>
<th>Idéal pour</th>
</tr>
</thead>
<tbody><tr>
<td>Portions avec la main</td>
<td>Gratuit</td>
<td>Faible, une fois maîtrisé</td>
<td>Un ordre de grandeur</td>
<td>Intuition, voyages</td>
</tr>
<tr>
<td>Objet de référence + table</td>
<td>Gratuit</td>
<td>Élevé</td>
<td>Valeurs détaillées par aliment</td>
<td>Vérifications ponctuelles</td>
</tr>
<tr>
<td>Assistant IA généraliste</td>
<td>Souvent gratuit</td>
<td>Faible</td>
<td>Estimation expliquée, sans journal</td>
<td>Questions ponctuelles</td>
</tr>
<tr>
<td>Application de suivi par photo</td>
<td>Téléchargement gratuit, formules payantes variables</td>
<td>Faible</td>
<td>Estimation modifiable, journal et historique</td>
<td>Une habitude quotidienne</td>
</tr>
</tbody></table>
<h2>Pour conclure</h2>
<p>Pas besoin de balance pour avoir une idée raisonnable de ce que vous mangez. Les portions avec la main sont un bon entraînement pour l'œil. Pour une habitude quotidienne, choisissez l'application photo que vous continuerez vraiment à ouvrir. Et quel que soit l'outil, lisez les chiffres comme des estimations : utiles pour repérer des tendances, pas un verdict sur chaque assiette.</p>
<p>Si vous voulez essayer Mealumo sur iPhone (pour rappel : l'app est en anglais et nécessite iOS 26+) : <a href="https://apps.apple.com/app/id6811904664">https://apps.apple.com/app/id6811904664</a>. Laissez un commentaire ou envoyez-moi un message privé et je vous enverrai un code promo PRO gratuit. Et si l'estimation de l'un de vos plats est vraiment à côté, je serai ravi de le savoir.</p>
]]></content:encoded></item><item><title><![CDATA[Kalorien einer Mahlzeit per Foto schätzen: 4 Methoden im Vergleich]]></title><description><![CDATA[Von der eigenen Hand bis zur spezialisierten App: was jede Methode kann, wo sie an Grenzen stößt und wie du eine findest, bei der du bleibst]]></description><link>https://mealumo.hashnode.dev/kalorien-einer-mahlzeit-per-foto-sch-tzen-4-methoden-im-vergleich</link><guid isPermaLink="true">https://mealumo.hashnode.dev/kalorien-einer-mahlzeit-per-foto-sch-tzen-4-methoden-im-vergleich</guid><category><![CDATA[german]]></category><category><![CDATA[Nutrition]]></category><category><![CDATA[Health,]]></category><category><![CDATA[AI]]></category><category><![CDATA[apps]]></category><dc:creator><![CDATA[Vlad]]></dc:creator><pubDate>Mon, 19 Oct 2026 10:26:56 GMT</pubDate><content:encoded><![CDATA[<p><em>Hinweis in eigener Sache: Ich bin der Entwickler von Mealumo, einer der unten genannten Apps. Der Ratgeber soll dir unabhängig davon helfen, wofür du dich entscheidest, auch bei den kostenlosen Methoden ganz ohne App. Wichtig: Mealumo ist nur auf Englisch verfügbar und setzt iOS 26 oder neuer voraus.</em></p>
<p><em>Dieser Beitrag wurde mit KI-Unterstützung verfasst und vor der Veröffentlichung geprüft.</em></p>
<p>Der Teller steht vor dir, du hättest gern eine grobe Kalorienzahl, aber keine Küchenwaage in Reichweite. Vielleicht sitzt du im Restaurant, vielleicht hast du einfach keine Lust, die Nudeln abzuwiegen. Hier sind vier praktische Wege, den Teller per Augenmaß einzuschätzen, von „ganz ohne Hilfsmittel“ bis zur eigenen App, jeweils mit Stärken und Schwächen.</p>
<p><strong>Vorab.</strong> Alle vier Methoden liefern eine Schätzung, keine Messung. Weder Mensch noch Software sieht, was in einem Auflauf steckt oder unter der Soße liegt. Für einen allgemeinen Überblick reicht das. Wenn du aus medizinischen Gründen genaue Werte brauchst, wiege dein Essen ab und lass dich von einer Ernährungsfachkraft oder deiner Ärztin bzw. deinem Arzt beraten.</p>
<h2>Methode 1: Portionen mit der Hand (kostenlos, ohne Handy)</h2>
<p>Deine Hand ist ein Messwerkzeug, das du immer dabeihast, und sie passt grob zu deiner Körpergröße. Eine verbreitete Faustregel:</p>
<ul>
<li><strong>Handfläche</strong> (Größe und Dicke): eine Portion Eiweiß, etwa Fleisch, Fisch, Tofu oder Eier</li>
<li><strong>Faust</strong>: eine Portion Gemüse</li>
<li><strong>Hohle Hand</strong>: eine Portion Kohlenhydrate, etwa Reis, Nudeln, Getreide oder Obst</li>
<li><strong>Daumen</strong>: eine Portion Fett, etwa Öl, Butter, Erdnussmus oder Käse</li>
</ul>
<p>Zähl die Portionen auf dem Teller, multipliziere jede mit einem typischen Kalorienwert der jeweiligen Gruppe, und du hast eine Hausnummer.</p>
<p><strong>Gut für:</strong> ein Gefühl für Mengen entwickeln, auf Reisen und überall dort, wo das Handy am Tisch fehl am Platz wäre. <strong>Achtung:</strong> Das Ergebnis ist grob, es braucht Übung, und „Handflächen“ werden gern unbemerkt großzügiger.</p>
<h2>Methode 2: Referenzobjekt plus Nährwertdatenbank (kostenlos)</h2>
<p>Fotografiere das Essen zusammen mit etwas, dessen Größe du kennst: einer Gabel, einer Getränkedose oder einer Scheckkarte. Damit schätzt du die Menge jeder Zutat und schlägst sie anschließend in einer Nährwertdatenbank nach. Kostenlos ist zum Beispiel USDA FoodData Central vom US-Landwirtschaftsministerium.</p>
<p><strong>Gut für:</strong> gelegentliche, sorgfältige Kontrollen mit verlässlichen Referenzdaten. <strong>Achtung:</strong> Pro Mahlzeit dauert das mehrere Minuten, deshalb hält man es selten täglich durch.</p>
<h2>Methode 3: Ein allgemeiner KI-Assistent (oft kostenlos)</h2>
<p>Mehrere KI-Assistenten, die Bilder verstehen, können ein Essensfoto beschreiben und die Nährwerte schätzen, wenn du sie darum bittest. Eine Anfrage, die gut funktioniert: <em>„Liste jedes Lebensmittel auf, das du siehst, schätze jeweils die Menge, nenne Kalorien, Eiweiß, Kohlenhydrate und Fett pro Posten und erkläre deine Annahmen.“</em></p>
<p><strong>Gut für:</strong> einzelne Fragen und um nachzuvollziehen, wie eine Zahl zustande kommt. <strong>Achtung:</strong> Es gibt kein Tagebuch, keine Ziele und keinen Verlauf. Die Antwort kann beim zweiten Fragen anders ausfallen, und die Zahlen musst du selbst irgendwo festhalten.</p>
<h2>Methode 4: Eine spezialisierte Foto-Tracking-App</h2>
<p>Solche Apps verbinden die Schätzung aus dem Foto mit einem Ernährungstagebuch, Tageszielen und einem Verlauf. Einige Beispiele laut ihren App-Store-Beschreibungen, Stand Oktober 2026 (Funktionen und Tarife ändern sich, also prüf vor der Entscheidung die aktuelle Seite):</p>
<ul>
<li><strong>MyFitnessPal:</strong> ein etablierter Kalorienzähler. Der Meal Scan ist als Premium-Funktion aufgeführt, zusammen mit Barcode-Scan und Spracheingabe.</li>
<li><strong>Lose It!:</strong> ebenfalls seit Langem auf dem Markt. Das Erfassen von Mahlzeiten per Foto gehört zu den Funktionen des Premium-Tarifs.</li>
<li><strong>Cal AI:</strong> ganz darauf ausgelegt, die Mahlzeit zu fotografieren. Laut Beschreibung setzen die Ergebnisse der Essensanalyse ein Abo voraus.</li>
<li><strong>SnapCalorie:</strong> Erfassung per Foto oder Sprachnotiz. Die App schätzt Lebensmittel und Portionen.</li>
<li><strong>Mealumo</strong> (meine App): nur iPhone, ab iOS 26, <strong>App-Sprache Englisch</strong>. Du machst ein Foto oder wählst eines aus der Mediathek. Die App schätzt Zutaten, Portionen, Kalorien, Eiweiß, Kohlenhydrate und Fett, und du prüfst und bearbeitest das Ergebnis vor dem Speichern. Dazu kommen Ernährungstagebuch, Lebensmittelsuche, Wasser- und Gewichtsprotokoll sowie Rezeptideen. Der Download ist kostenlos, ein optionales PRO-Abo schaltet weiteres Scannen frei. Die Fotoanalyse läuft über einen externen KI-Dienst und braucht eine Internetverbindung.</li>
</ul>
<p><strong>Gut für:</strong> eine tägliche Routine. Ein Foto genügt, und mit dem wachsenden Verlauf werden Muster sichtbar. <strong>Achtung:</strong> kostenpflichtige Tarife, eine Genauigkeit, die von Gericht zu Gericht schwankt, und alles, was die Kamera nicht erfasst, musst du weiterhin selbst ergänzen.</p>
<p><strong>So wählst du aus:</strong> Teste ein paar Apps einige Tage lang mit deinen ganz normalen Mahlzeiten. Nimm die, deren Ergebnisse du am leichtesten <em>überprüfen und korrigieren</em> kannst, nicht die mit dem lautesten Marketing.</p>
<h2>Bessere Schätzungen, egal mit welcher Methode</h2>
<ul>
<li><strong>Lass Platz um den Teller.</strong> Ein knapp beschnittenes Foto schneidet Beilagen, Soße und das Glas daneben ab.</li>
<li><strong>Gutes Licht schlägt schönen Winkel.</strong> Schatten und gedimmtes Restaurantlicht lassen Lebensmittel ähnlich aussehen.</li>
<li><strong>Ergänze, was du weißt.</strong> Wenn du selbst gekocht hast, weißt du, ob es Vollmilch oder fettarme Milch war und wie viel Butter drin ist. Trag es nach.</li>
<li><strong>Erst das Gericht, dann die Extras.</strong> Erfasse zuerst den Hauptteller und dann alles, was dazukam: Dressing, Soße, Zucker im Kaffee.</li>
<li><strong>Bleib bei einer Methode.</strong> Wer alle paar Tage wechselt, kann seine Zahlen kaum noch vergleichen.</li>
</ul>
<h2>Auf einen Blick</h2>
<table>
<thead>
<tr>
<th>Methode</th>
<th>Kosten</th>
<th>Aufwand pro Mahlzeit</th>
<th>Ergebnis</th>
<th>Am besten für</th>
</tr>
</thead>
<tbody><tr>
<td>Portionen mit der Hand</td>
<td>Kostenlos</td>
<td>Gering, wenn geübt</td>
<td>Grobe Hausnummer</td>
<td>Mengengefühl, Reisen</td>
</tr>
<tr>
<td>Referenzobjekt + Datenbank</td>
<td>Kostenlos</td>
<td>Hoch</td>
<td>Detaillierte Werte pro Lebensmittel</td>
<td>Gelegentliche Kontrollen</td>
</tr>
<tr>
<td>Allgemeiner KI-Assistent</td>
<td>Oft kostenlos</td>
<td>Gering</td>
<td>Erklärte Schätzung, kein Tagebuch</td>
<td>Einzelne Fragen</td>
</tr>
<tr>
<td>Foto-Tracking-App</td>
<td>Gratis-Download, Bezahltarife unterschiedlich</td>
<td>Gering</td>
<td>Bearbeitbare Schätzung, Tagebuch und Verlauf</td>
<td>Tägliche Routine</td>
</tr>
</tbody></table>
<h2>Fazit</h2>
<p>Für eine vernünftige Vorstellung davon, was du isst, brauchst du keine Waage. Mit Handportionen schulst du dein Auge. Für die tägliche Routine nimm die Foto-App, die du auch wirklich weiter öffnest. Und egal womit: Lies die Zahlen als Schätzungen, hilfreich, um Muster zu erkennen, aber kein Urteil über einen einzelnen Teller.</p>
<p>Wenn du Mealumo auf dem iPhone ausprobieren möchtest (zur Erinnerung: die App ist auf Englisch und braucht iOS 26+): <a href="https://apps.apple.com/app/id6811904664">https://apps.apple.com/app/id6811904664</a>. Schreib mir einen Kommentar oder eine Direktnachricht, dann schicke ich dir einen kostenlosen PRO-Promo-Code. Und wenn die Schätzung für eins deiner Gerichte deutlich danebenliegt, würde ich das gern wissen.</p>
]]></content:encoded></item><item><title><![CDATA[Building a photo-to-calories logger for iOS: the AI can't see the olive oil]]></title><description><![CDATA[Turning a meal photo into an editable food entry, and why the review step matters more than the model Disclosure: I'm the developer of Mealumo, the iOS app…]]></description><link>https://mealumo.hashnode.dev/building-a-photo-to-calories-logger-for-ios-the-ai-can-t-see-the-olive-oil</link><guid isPermaLink="true">https://mealumo.hashnode.dev/building-a-photo-to-calories-logger-for-ios-the-ai-can-t-see-the-olive-oil</guid><category><![CDATA[iOS]]></category><category><![CDATA[AI]]></category><category><![CDATA[UX]]></category><category><![CDATA[Swift]]></category><category><![CDATA[Mobile Development]]></category><dc:creator><![CDATA[Vlad]]></dc:creator><pubDate>Mon, 19 Oct 2026 09:53:39 GMT</pubDate><content:encoded><![CDATA[<p><em>Turning a meal photo into an editable food entry, and why the review step matters more than the model</em></p>
<p><strong>Disclosure:</strong> I'm the developer of Mealumo, the iOS app this post is about. This is a write-up of the design decisions behind it, not a review. The link is at the end. If you think we got a tradeoff wrong, I'd like to hear it in the comments.</p>
<p><em>This post was drafted with AI assistance and reviewed before publishing.</em></p>
<hr />
<h2>The problem</h2>
<p>When people give up on calorie trackers, it's usually not the nutrition that beats them. It's the data entry. You type "chicken salad", scroll through dozens of near-identical entries, guess a portion, then do it all again for the dressing, the bread and the drink. Doing that three to five times a day is a lot to ask.</p>
<p>So the idea behind Mealumo is simple: let a photo start the entry. You take a picture of your plate, or pick one from your photo library, and the app proposes ingredients, portions, calories, protein, carbs and fat. Then you fix whatever is wrong and save it.</p>
<p>That last sentence turned out to be the real product.</p>
<h2>The flow</h2>
<p>Here's how an entry gets made:</p>
<ol>
<li><p><strong>Capture.</strong> Use the camera or choose an existing photo. Library import means a meal photographed at lunch can be logged in the evening.</p>
</li>
<li><p><strong>Analyze.</strong> The photo goes to an external AI service, which estimates what's on the plate: individual ingredients, their portions, and the calories and macros for each. This step needs an internet connection.</p>
</li>
<li><p><strong>Review.</strong> The estimate is shown as a list of ingredients you can edit, not as one final number. You adjust portions, edit ingredients, and add anything that's missing.</p>
</li>
<li><p><strong>Save.</strong> The entry goes into a daily journal, which adds up calories and macros against your daily targets.</p>
</li>
</ol>
<p>Around that core sit the less glamorous parts: food search for adding ingredients by hand, a water log, a weight log with progress over time, recipe ideas, and daily nutrition targets estimated from your goals, body measurements and activity.</p>
<h2>Why one number is the wrong output</h2>
<p>The tempting version of this feature is the demo version: photo in, "640 kcal" out. It looks great, but it fails in a way the user can't fix.</p>
<p>Suppose the app says 640 and you know the portion was bigger. What do you edit? If you overwrite the total, the macros stop adding up, and you still don't know which part of the meal was wrong.</p>
<p>Breaking the estimate into <strong>ingredients Ã— portions</strong> fixes that. Most mistakes on a food photo are local: the amount of rice is misjudged, or a sauce turns out to be cream-based rather than tomato. When the output is a list, you correct the one wrong line and the totals follow. The model doesn't need to be perfect. It needs to be a better starting point than an empty search box.</p>
<p>An itemized estimate also has practical benefits on the app side. Each line can be shown, checked and edited on its own, which is much easier than parsing a paragraph of free text.</p>
<h2>What a photo physically can't tell you</h2>
<p>The model isn't the hard limit. Some information just isn't in the pixels:</p>
<ul>
<li><p><strong>Fats and oils.</strong> A tablespoon of olive oil is roughly 120 kcal and becomes basically invisible once it's cooked into the food. The same goes for butter on vegetables and oil in a dressing.</p>
</li>
<li><p><strong>Portion depth.</strong> A photo taken from above shows area, not volume. A deep bowl and a shallow plate can look the same.</p>
</li>
<li><p><strong>Hidden layers.</strong> Burritos, sandwiches, casseroles: anything mixed or wrapped.</p>
</li>
<li><p><strong>Recipes.</strong> "Pasta with sauce" might be a light tomato sauce or a heavy cream one.</p>
</li>
<li><p><strong>Drinks.</strong> Two cups can look alike and contain very different things.</p>
</li>
</ul>
<p>You can't engineer these away. What you can do is be honest about them and make corrections cheap. That's why the App Store listing says outright that nutrition estimates may be inaccurate and should be reviewed before use, and why reviewing the estimate is part of the flow, not an optional extra.</p>
<p>An app that users trust a little less, so they check for the oil, is better than one they trust completely until the numbers stop making sense.</p>
<h2>Speed vs. trust</h2>
<p>There's a real tension here. Logging only works if it's quicker than typing, and every confirmation step adds friction back.</p>
<p>These are the principles we designed around:</p>
<ul>
<li><p><strong>Show the estimate right away, and make it editable.</strong> The first thing you see after a scan should be something you can act on.</p>
</li>
<li><p><strong>Treat portions as the main thing to correct.</strong> Portion size is what a photo shows least reliably, so adjusting it has to be easy.</p>
</li>
<li><p><strong>Keep manual entry close by.</strong> Some meals don't photograph well, like a shake in an opaque bottle or a meal that's already half eaten. Food search covers those.</p>
</li>
<li><p><strong>Never claim more precision than exists.</strong> An estimate should look like an estimate.</p>
</li>
</ul>
<h2>Privacy is a product decision, not a footnote</h2>
<p>Meal photos are personal. Sending them to a third-party service is a real decision, and users should know about it. The Mealumo listing says plainly that meal photos, and the profile details needed to generate nutrition targets, are sent to an external AI service for processing. Putting that sentence on the store page, rather than burying it in a policy, is a small thing. It's still the kind of thing that decides whether people trust a health-adjacent app.</p>
<h2>Not a medical app</h2>
<p>A calorie tracker can easily slide into health claims. Mealumo is a general-wellness food journal. Its targets are estimates, it doesn't diagnose or treat anything, and the listing says so. It doesn't promise weight loss, and I don't think a tracker should. What it can do is make the tedious part of paying attention to food take less time.</p>
<h2>Takeaways for anyone building on vision models</h2>
<ul>
<li><p><strong>Return structure, not a verdict.</strong> Itemized output lets users fix one item instead of throwing the whole result away.</p>
</li>
<li><p><strong>Know what your input can't contain.</strong> For food it's oil, depth and hidden layers. Every domain has its own version of this.</p>
</li>
<li><p><strong>Put the uncertainty in the UI.</strong> A disclaimer helps, but an editable result communicates "this is a guess" better than any fine print.</p>
</li>
<li><p><strong>Be explicit about where the data goes.</strong> Users notice.</p>
</li>
</ul>
<h2>Try it / tell me where it's wrong</h2>
<p>Mealumo is free to download on iPhone (iOS 26+), with an optional PRO subscription for continued scanning and the full recipe collection: <a href="https://apps.apple.com/app/id6811904664">https://apps.apple.com/app/id6811904664</a></p>
<p>It's new and has no ratings yet, so feedback counts for a lot right now, especially photos where the estimate is clearly off. If you want to try PRO, comment or DM me for a free PRO promo code.</p>
<p>And if you've built something on vision models with messy real-world input, how did you handle the cases where the answer simply isn't in the image?</p>
]]></content:encoded></item><item><title><![CDATA[How to estimate calories from a photo of your meal: 4 methods compared]]></title><description><![CDATA[From your own hand to a dedicated app, with what each method is good at, where it falls short, and how to pick one you'll stick with]]></description><link>https://mealumo.hashnode.dev/how-to-estimate-calories-from-a-photo-of-your-meal-4-methods-compared</link><guid isPermaLink="true">https://mealumo.hashnode.dev/how-to-estimate-calories-from-a-photo-of-your-meal-4-methods-compared</guid><category><![CDATA[Nutrition]]></category><category><![CDATA[Health,]]></category><category><![CDATA[AI]]></category><category><![CDATA[Mobile apps]]></category><category><![CDATA[guide]]></category><dc:creator><![CDATA[Vlad]]></dc:creator><pubDate>Mon, 19 Oct 2026 09:43:58 GMT</pubDate><content:encoded><![CDATA[<p><em>Disclosure: I'm the developer of Mealumo, one of the apps mentioned below. I've written this so it's useful whichever option you choose, including the free ones that don't involve an app at all.</em></p>
<p>You have a plate in front of you and want a rough calorie number, but there's no kitchen scale nearby. You might be at a restaurant, or you just don't feel like weighing rice. This guide covers four practical ways to estimate calories from what you can see, from no tools at all to a dedicated app, with the tradeoffs of each.</p>
<p><strong>One thing first.</strong> All four methods give you an estimate, not a measurement. Judging food by eye is limited whether a person or software does it: what's inside a dish or under the surface simply isn't visible. That's fine if you want general awareness. If you need exact numbers for medical reasons, weigh your food and work with a registered dietitian or your doctor.</p>
<h2>Method 1: Hand portions (free, no phone)</h2>
<p>Your hand is a measuring tool you always have with you, and it scales roughly with your body size. A widely used rule of thumb:</p>
<ul>
<li><strong>Palm</strong> (size and thickness): one portion of protein, such as meat, fish, tofu or eggs</li>
<li><strong>Fist</strong>: one portion of vegetables</li>
<li><strong>Cupped hand</strong>: one portion of carbohydrates, such as rice, pasta, grains or fruit</li>
<li><strong>Thumb</strong>: one portion of fats, such as oil, butter, nut butter or cheese</li>
</ul>
<p>Count the portions on your plate and multiply each by a typical calorie value for that food group. That gives you a ballpark figure.</p>
<p><strong>Good for:</strong> building intuition, travel, and situations where pulling out a phone feels awkward. <strong>Watch out for:</strong> it's coarse, it takes practice, and "palms" have a way of getting generous.</p>
<h2>Method 2: A reference object plus a nutrition database (free)</h2>
<p>Take the photo with something of known size next to the food, like a fork, a standard drink can or a credit card. Use it to judge how much of each item there is, then look each item up in a nutrition database. USDA FoodData Central is free and maintained by the U.S. Department of Agriculture.</p>
<p><strong>Good for:</strong> occasional careful checks with trustworthy reference data. <strong>Watch out for:</strong> it takes several minutes per meal, which makes it hard to keep up every day.</p>
<h2>Method 3: A general-purpose AI assistant (often free)</h2>
<p>Several multimodal AI assistants can describe a food photo and estimate its nutrition if you ask. A prompt that works well: <em>"List each food you see, estimate the portion of each, give calories, protein, carbs and fat per item, and state your assumptions."</em></p>
<p><strong>Good for:</strong> one-off questions, and seeing the reasoning behind a number. <strong>Watch out for:</strong> there's no food log, no goals and no history, the answer can change if you ask twice, and you have to record the numbers somewhere else yourself.</p>
<h2>Method 4: A dedicated photo-based tracking app</h2>
<p>These apps combine the photo estimate with a food diary, daily targets and a record over time. A few examples, based on their App Store descriptions as of October 2026. Features and plans change, so check the current listing before you decide:</p>
<ul>
<li><strong>MyFitnessPal:</strong> a long-established tracker. Meal scan is listed as a Premium feature, alongside barcode scan and voice logging.</li>
<li><strong>Lose It!:</strong> another long-running tracker. Photo meal logging is listed among its Premium plan features.</li>
<li><strong>Cal AI:</strong> built around snapping a photo of a meal. Its listing notes that food scanning results require a subscription.</li>
<li><strong>SnapCalorie:</strong> log by photo or voice note, and the app estimates the food and portions.</li>
<li><strong>Mealumo</strong> (the one I develop): iPhone only, iOS 26+. Take a photo or pick one from your library, and it estimates ingredients, portions, calories, protein, carbs and fat, which you review and edit before saving. It also has a daily food journal, food search, water and weight logs, and recipe ideas. It's free to download, with an optional PRO subscription for continued scanning. Photo analysis uses an external AI service and needs an internet connection.</li>
</ul>
<p><strong>Good for:</strong> a daily habit, since logging takes one photo and your history builds up so patterns become visible. <strong>Watch out for:</strong> paid tiers, accuracy that varies from dish to dish, and the fact that you still have to add what the camera missed.</p>
<p><strong>How to choose:</strong> try a few of them on your own ordinary meals for several days. Go with the one whose results you find easiest to <em>check and fix</em>, not the one with the boldest marketing.</p>
<h2>Getting a better estimate, whichever method you use</h2>
<ul>
<li><strong>Leave space around the plate.</strong> A cropped photo hides sides, sauces and the glass next to it.</li>
<li><strong>Choose good light over a nice angle.</strong> Shadows and dim restaurant lighting make foods harder to tell apart.</li>
<li><strong>Say what you know.</strong> If you made it yourself, you know whether it was whole milk or skim, and how much butter went in. Add that.</li>
<li><strong>Log the item, then the extras.</strong> Note the main dish first, then go back for anything added afterwards, like dressing, a sauce or sugar in your coffee.</li>
<li><strong>Use one method consistently.</strong> Switching methods every few days makes your numbers harder to compare.</li>
</ul>
<h2>Side-by-side</h2>
<table>
<thead>
<tr>
<th>Method</th>
<th>Cost</th>
<th>Effort per meal</th>
<th>What you get</th>
<th>Best for</th>
</tr>
</thead>
<tbody><tr>
<td>Hand portions</td>
<td>Free</td>
<td>Low, once learned</td>
<td>A rough ballpark</td>
<td>Intuition, travel</td>
</tr>
<tr>
<td>Reference object + database</td>
<td>Free</td>
<td>High</td>
<td>Itemized, database-backed numbers</td>
<td>Occasional careful checks</td>
</tr>
<tr>
<td>General AI assistant</td>
<td>Often free</td>
<td>Low</td>
<td>An explained estimate, but no log</td>
<td>One-off questions</td>
</tr>
<tr>
<td>Dedicated photo app</td>
<td>Free download, paid tiers vary</td>
<td>Low</td>
<td>Editable estimate plus diary and history</td>
<td>A daily habit</td>
</tr>
</tbody></table>
<h2>Bottom line</h2>
<p>You don't need a scale to get a reasonable sense of what you eat. Hand portions are a good way to build intuition. For a daily habit, pick the photo app you'll actually keep opening. Whatever you use, read the numbers as estimates: useful for spotting patterns, not a verdict on any single plate.</p>
<p>If you want to try Mealumo on iPhone: <a href="https://apps.apple.com/app/id6811904664">https://apps.apple.com/app/id6811904664</a>. Comment or DM me for a free PRO promo code. And if its estimate for one of your meals is clearly off, I'd like to know.</p>
]]></content:encoded></item></channel></rss>