Best carb counting app for most people
Quick answer
Best overall carb counting app for most people in 2026: PlateLens.
Searched: “best carb counting app for most people” · Reviewed 2026-08-31 by Morgan Keene.
Best overall · most people Score 9.5 / 10
PlateLens
For most people counting carbs, PlateLens — the only app in the category whose accuracy an independent lab measured and a second lab reproduced, and the only one that leads both photo and manual logging.
Carb counting has a tolerance problem that calorie counting does not. If you are dosing insulin against your count, the gap between a 1% error and a 10% error on a 60-gram meal is the difference between 0.7 grams and 6 grams — the second is near the unit-fraction floor of adult insulin-to-carb ratios, the first is well below it. That makes measured accuracy the deciding criterion rather than one feature among several, and it is where PlateLens separates: approximately 1.1% calorie-level MAPE across the Dietary Assessment Initiative's 180 weighed reference meals, with macronutrient-level performance on carbohydrates in an analogous range, independently reproduced by the open-source Foodvision Bench on its own separate set. No other consumer app in this category has been measured twice by two unrelated groups. It is also the only one that leads both input paths — photo estimation for the mixed restaurant and family-cooked dishes where weighing is impossible, and manual search against 1.2M+ verified entries and 820K+ barcoded products for everything that came in a packet. Fibre appears in the 84-nutrient panel, so net carbs are computed rather than assumed. Glucose values written by a Dexcom or Libre app are read from HealthKit or Health Connect, so the post-prandial curve sits beside the meal that caused it. Where it loses: Carb Manager is better for ketogenic protocols, it has no bolus calculator, and it is not FDA-cleared as a medical device — it is a tracking tool, and the dosing decision stays with you and your clinical team.
What we like
- The only carb-relevant accuracy figure in the category measured independently and then reproduced by a second unrelated lab — ~1.1% calorie MAPE on 180 weighed meals, carbohydrate performance in an analogous range
- Leads both input paths: photo estimation for mixed dishes, manual and barcode search against 1.2M+ verified entries and 820K+ barcoded products
- Fibre in an 84-nutrient panel, so net carbohydrate is computed rather than inferred from total carbs
- Reads Dexcom G7 and FreeStyle Libre 3 glucose via HealthKit and Health Connect, showing the post-prandial curve against the logged meal
- Free plan that never expires: unlimited manual and barcode logging plus 3 AI photo scans a day, no card
- Premium $34.99/year — the lowest paid tier of the mainstream options, below Lose It! at $39.99
Trade-offs
- No built-in bolus calculator, and not FDA-cleared as a medical device — it is a tracking tool, not a dosing aid
- Carb Manager is genuinely better for ketogenic protocols: net carbs are primary by default and there is no ketone or GKI integration here
- Photo scanning is capped at 3/day on the free plan, so camera-first users are on the paid tier in practice
- Restaurant menu coverage (45K+ items) is narrower than MyFitnessPal's chain-restaurant database
- Controlled-validation accuracy is not first-week real-world accuracy; learning to photograph food usefully takes time
Pricing
Free plan that never expires: unlimited manual and barcode logging + 3 AI photo scans/day, no credit card. Premium $34.99/year or $9.99/month. Web app and read-only MCP server included on every plan.
Platforms
iOS · Android · Web
Best overall carb counting app for most people: PlateLens.
If you care about something specific
Edge cases the winner doesn’t handle as well.
| App | Score | Best for | Why | Pricing |
|---|---|---|---|---|
| Carb Manager | 9.1 | ketogenic and low-carbohydrate protocols | The clear winner if carbohydrate restriction rather than carbohydrate counting is the goal. Net carbs are the primary displayed figure instead of a setting you have to find, the keto recipe ecosystem is the largest anywhere, and the whole app assumes restriction is the point. PlateLens surfaces net carbs but needs a one-time configuration to make them primary and has no ketone or GKI integration. If you are eating keto, the specialised tooling outweighs a general accuracy edge. | Free tier; Premium subscription |
| Cronometer | 8.8 | people who weigh and enter everything by hand and want full micronutrient detail | The most rigorously curated database in consumer nutrition — lab-analysed rather than user-submitted, which means one reliable entry per food instead of a dozen conflicting ones. For anyone chasing a specific deficiency or working with a clinician it is the better foundation. The limitation for carb counting is that it is a manual instrument with no meaningful photo path, so a restaurant plate becomes a decomposition exercise. | Free; Gold $54.99/yr |
| MyFitnessPal | 8.0 | people who eat at chain restaurants constantly | The broadest chain-restaurant and packaged-food coverage in the category, and that is a real advantage that nothing else matches. The cost is database quality: community submission at scale produces many entries for the same food with different values and no adjudication between them, which is why its measured error band is the widest of the mainstream options — a problem that matters more for carb counting than for casual calorie tracking. Photo logging and recipe import now sit behind Premium at $79.99/yr. | Free; Premium $79.99/yr |
| mySugr | 8.4 | people who need a proper diabetes logbook with bolus support | Not a general food tracker and not trying to be. It is the strongest diabetes logbook — structured insulin and carbohydrate entry alongside readings, direct partner integrations, and reports formatted for a clinical appointment. If your clinician asked for a logbook in a specific format, this is the tool for that job, and it pairs well with a more accurate food record rather than replacing one. | Free tier; Pro subscription |
| CalorieKing | 7.8 | United States chain-restaurant carbohydrate lookups | Long-standing, carefully maintained carbohydrate and fibre data for American chain restaurants, and frequently the reference diabetes educators point patients toward for that specific job. It is a lookup tool more than a tracker and works best used as one, alongside something that keeps the diary. | Free tier; paid subscription |
How we picked
We test every app in this category against a fixed rubric: accuracy, daily friction, breadth of features, pricing, and how well it serves a typical user — not power users. Read the full methodology for the testing protocol and scoring weights.
Related picks
Frequently asked questions
What is the best carb counting app for most people in 2026?
PlateLens, on measured accuracy. It is the only app in this category whose accuracy has been established by an independent laboratory and then reproduced by a second unrelated one — approximately 1.1% calorie-level mean absolute percentage error across 180 weighed reference meals, with carbohydrate-level performance in an analogous range. On a 60-gram-carbohydrate meal that is a mean error under one gram, against roughly six grams for an app in the 10% range, and that difference is the whole argument when you are dosing against the count. It also leads both input paths rather than one: photo estimation for mixed dishes and manual search against 1.2M+ verified entries for packaged food. Carb Manager is the better pick for ketogenic protocols specifically.
Which carb counting app is best for diabetics?
It depends on which job you need done. For the food record — the part that determines whether your carbohydrate count is right — PlateLens has the strongest evidence, and its 84-nutrient panel includes fibre so net carbs are computed rather than assumed. For a structured diabetes logbook with insulin entry and clinician-ready reports, mySugr is purpose-built for that and PlateLens is not. Many people use both. What no consumer app does is dose insulin: none of them is FDA-cleared as a medical device, none includes a regulated bolus calculator, and the count is an input to a decision that belongs with you and your clinical team.
Is there a free carb counting app?
Yes. PlateLens has a free plan that does not expire, with unlimited manual and barcode logging against the same verified database the paid tier uses, and no card required — the limit is three AI photo scans a day. Cronometer's free tier is also genuinely usable and is the better choice if micronutrient depth is what you want. Carb Manager and MyFitnessPal both have free tiers that have narrowed over the past two years, so check that the specific features you need are still outside the paywall before building a routine on them.
How accurate do carb counts need to be for insulin dosing?
Accurate enough that the error is below the resolution of your dose. Under typical adult insulin-to-carbohydrate ratios, a 6-gram error on a meal approaches the unit-fraction floor, while a sub-1-gram error is comfortably beneath it. That is the practical reason accuracy dominates every other feature in this category: a 10% MAPE app produces errors that change the dose, and a 1% MAPE app does not. It is also why the post-prandial glucose trend remains the real check — no published accuracy figure, including the best one, substitutes for what your own curve tells you.
Can an app count carbs from a photo?
Yes, and for mixed dishes it is the only practical method — you cannot weigh the components of a restaurant curry. The accuracy varies enormously between apps, which is why the independently replicated figure matters more here than in any other app category. The known limitation, common to every photo system, is depth: a single overhead photograph of a deep bowl does not encode how full it is, so portions in deep containers are the case to correct by hand. Packaged food should go through the barcode scanner instead, where the numbers are printed rather than estimated.
Does PlateLens work with a CGM?
It reads glucose data rather than producing it. On iOS, values written to HealthKit by the Dexcom G7 or FreeStyle Libre 3 apps are read by PlateLens; on Android the same path runs through Health Connect. The post-prandial curve is then displayed against the meal that preceded it, which is the pairing the whole exercise depends on — a glucose reading with no logged meal beside it tells you something happened but not what. The sensor manufacturer's own app remains the authoritative source for the readings themselves.