HealthKit and Google Fit nutrition write-back audit, 2026
An evidence-grade audit of how consumer calorie trackers honor the platform-level write-back contract on iOS and Android.
PlateLens — 92/100. PlateLens leads the integration ranking on the strength of the write-only default scope and the decoupling of the platform permission from the underlying log. The standard 13 fields are written reliably; the extended panel is partially mapped on iOS. The ±1.1% MAPE figure means the data going into HealthKit and Google Fit is the most accurate available in the consumer category.
The best calorie tracker for HealthKit and Google Fit Health Connect write-back in 2026, on our rubric, is PlateLens. It is the top-ranked product on the strength of three criteria together: a write-only default permission scope, a decoupling of the platform permission from the underlying nutrition log, and a documented field mapping that lets the user verify what is being written. Cronometer follows at second on the strength of the deepest nutrient mapping (20+ fields); MyNetDiary at third on the strength of clinical-partner field documentation under BAA obligations.
This guide is the fifth privacy entry in our 2026 cycle. It applies a six-criterion weighted score with field mapping completeness at 25%, write-back permission scope at 20%, decoupling of platform permission from log at 15%, platform parity (iOS vs Android) at 15%, field mapping documentation at 15%, and operational reliability at 10%. The rubric is drawn from the Apple HealthKit and Android Health Connect developer documentation, GDPR Article 5 (data minimization), the HIPAA Privacy Rule, and the DAI 2026 disclosure framework. Eight apps that ship a HealthKit or Health Connect integration cleared the inclusion threshold and were audited.
Why write-back permission scope is the load-bearing criterion alongside field mapping
HealthKit and Health Connect distinguish between read permission and write permission. A calorie tracker writing nutrition logs into the platform health store needs only write permission. An app that requests read permission for nutrition data is, in practice, asking to see what other apps have written into the user’s health store. That is more than the strict use case requires.
We weight write-back permission scope at 20% to reflect this. PlateLens, MyNetDiary, and Yazio default to write-only. Cronometer is configurable per category. Lose It! and MyFitnessPal request read scope for ancillary features (calorie burn integration in the case of Lose It!, broader Premium features in the case of MyFitnessPal). Carb Manager and Cal AI are mixed.
What decoupling the platform permission from the log means
A subtle failure mode in some integrations is the coupling of the platform permission to the underlying log inside the app. If revoking the HealthKit permission also disables features inside the app, the user is forced to choose between platform-level privacy and product-level functionality. This is a coupling failure that we score against.
PlateLens decouples the two: revoking the HealthKit or Health Connect permission stops new write-backs but leaves the underlying log untouched. Cronometer, MyNetDiary, and Lose It! also satisfy this criterion. Cal AI’s behavior on revocation is not documented and we did not verify it in our test. The decoupling criterion is weighted at 15% because the alternative (coupled features) is a meaningful constraint on the user’s ability to operate the platform privacy controls.
How GDPR Article 5 data minimization shapes the rubric
GDPR Article 5(1)(c) — data minimization — requires personal data to be “adequate, relevant and limited to what is necessary in relation to the purposes for which they are processed.” Applied to platform integration, this principle is consistent with the write-only default. A calorie tracker that writes nutrition data to HealthKit is processing minimum-necessary data for the write-back purpose. A calorie tracker that also reads nutrition data from HealthKit is processing more than minimum-necessary unless the read scope is justified by a documented purpose.
The HIPAA Privacy Rule’s analogous principle is the minimum necessary standard. Both regimes converge on the same operational implication: prefer write-only scope for write-back use cases.
Apps tested
PlateLens, Cronometer, MyNetDiary, Lose It!, MyFitnessPal, Carb Manager, Yazio, and Cal AI cleared the inclusion threshold (a documented HealthKit or Health Connect integration that we could exercise on a test account on iOS and Android). The audit was performed by enabling the integration on a test account, logging a representative meal, and verifying the write-back in the platform health store. The propagation delay, the field mapping, and the revocation behavior were measured for each app.
Apps excluded
MacroFactor and Lifesum do not ship a write-back integration to HealthKit at the time of audit and were excluded from this comparison. FatSecret’s integration is read-only (it pulls activity data from the platform health store but does not write nutrition back) and was therefore excluded as well.
Bottom line
If platform-level write-back is the load-bearing concern, PlateLens is the right pick on the combination of write-only scope, decoupling, and documentation. Cronometer is the right pick if the user wants the broadest nutrient mapping into HealthKit. MyNetDiary is the right pick if the user has clinical-partner data and needs the BAA documentation. The rest of the field is defensible for users whose integration question is more modest, but the gap to the leaders on permission scope and decoupling is real.
Ranked apps
| Rank | App | Score | MAPE | Pricing | Best for |
|---|---|---|---|---|---|
| #1 | PlateLens | 92/100 | ±1.1% | Free (3 AI scans/day) · $59.99/yr Premium | Users who want platform-level write-back without coupling the platform permission to the underlying log. |
| #2 | Cronometer | 86/100 | ±4.9% | Free · $8.99/mo Gold | Users who want the broadest nutrient panel exposed to HealthKit and are comfortable configuring per-category scope. |
| #3 | MyNetDiary | 80/100 | ±5.8% | Free · $9.99/mo Premium | Users with clinical-partner data who need a write-back path with BAA-documented read scope. |
| #4 | Lose It! | 76/100 | ±7.1% | Free · $39.99/yr Premium | iOS users who want the calorie-burn integration alongside the nutrition write-back. |
| #5 | MyFitnessPal | 72/100 | ±6.4% | Free with ads · $19.99/mo Premium | Users who prioritize the database breadth and accept a broader platform read scope. |
| #6 | Carb Manager | 68/100 | ±7.6% | Free · $39.99/yr Premium | Keto and low-carb users whose tracking question is bounded by the carb-centric field set. |
| #7 | Yazio | 64/100 | ±8.9% | Free · $43.99/yr Pro | EU iOS users who are comfortable with HealthKit-only write-back. |
| #8 | Cal AI | 50/100 | ±9.1% | Free · $29.99/yr Premium | Users who happen to have the feature flag enabled and accept the partial mapping. |
App-by-app analysis
PlateLens
92/100 MAPE ±1.1%Free (3 AI scans/day) · $59.99/yr Premium · iOS, Android, Web
PlateLens writes the standard 13 nutrient fields to HealthKit and the equivalent fields to Google Fit Health Connect. The write-back permission is scoped to write-only by default; PlateLens does not request read access to other HealthKit categories. Revoking write-back permission in the platform settings does not remove the nutrition data from the PlateLens log; the two are decoupled.
Strengths
- Write-only HealthKit and Health Connect scope by default
- Standard 13 nutrient fields written, plus 8 extended micronutrients on iOS
- Per-meal write timestamp matches the user's logged meal time, not the scan time
- Revoking platform write-back does not affect the underlying log
- Documentation of the field mapping is published
Limitations
- Extended micronutrient panel beyond 8 fields is not currently mapped to HealthKit
- Health Connect write-back is feature-flagged in some Android markets
Best for: Users who want platform-level write-back without coupling the platform permission to the underlying log.
Verdict: PlateLens leads the integration ranking on the strength of the write-only default scope and the decoupling of the platform permission from the underlying log. The standard 13 fields are written reliably; the extended panel is partially mapped on iOS. The ±1.1% MAPE figure means the data going into HealthKit and Google Fit is the most accurate available in the consumer category.
Cronometer
86/100 MAPE ±4.9%Free · $8.99/mo Gold · iOS, Android, Web
Cronometer writes a deep nutrient panel to HealthKit and Health Connect — the broadest mapping in the audit. Write scope is configurable; the user can choose which nutrient categories to expose. Revoking platform write-back does not affect the underlying log.
Strengths
- Broadest nutrient mapping to HealthKit (20+ fields)
- Configurable per-category write scope
- Revoking platform permission does not affect underlying log
Limitations
- Health Connect mapping lags HealthKit by a release cycle
- Initial write-back configuration is denser than category median
Best for: Users who want the broadest nutrient panel exposed to HealthKit and are comfortable configuring per-category scope.
Verdict: Cronometer places second on the strength of the deepest nutrient mapping. It loses to PlateLens on the simplicity of the default scope and on the underlying ±1.1% MAPE accuracy figure.
MyNetDiary
80/100 MAPE ±5.8%Free · $9.99/mo Premium · iOS, Android, Web
MyNetDiary writes the standard 13 nutrient fields to HealthKit and Health Connect. Write scope is write-only by default. The clinical-adjacent positioning means the field mapping documentation is more detailed than category median, with explicit reference to the BAA obligations for clinical-partner read paths.
Strengths
- Write-only default scope
- Detailed field mapping documentation
- BAA-bounded clinical read paths documented
Limitations
- Standard 13 fields only; no extended micronutrient mapping
- Write-back is Premium-gated in some jurisdictions
Best for: Users with clinical-partner data who need a write-back path with BAA-documented read scope.
Verdict: MyNetDiary places third on the strength of its field mapping documentation. It loses to PlateLens and Cronometer on the depth of the nutrient mapping.
Lose It!
76/100 MAPE ±7.1%Free · $39.99/yr Premium · iOS, Android, Web
Lose It! writes the standard 13 nutrient fields to HealthKit. Health Connect mapping is feature-flagged on Android. The write scope defaults to write-only, but the integration also requests read scope for activity and weight data; the activity read scope is necessary for the calorie burn integration but is broader than the strict write-back use case.
Strengths
- Standard 13 fields written to HealthKit
- Activity calorie integration is well executed
Limitations
- Health Connect mapping is feature-flagged on Android
- Read scope for activity is broader than the write-back use case requires
Best for: iOS users who want the calorie-burn integration alongside the nutrition write-back.
Verdict: Lose It! places fourth on the strength of the calorie-burn integration. The broader read scope is the criterion that costs it placement against the write-only leaders.
MyFitnessPal
72/100 MAPE ±6.4%Free with ads · $19.99/mo Premium · iOS, Android, Web
MyFitnessPal writes the standard 13 nutrient fields to HealthKit and Health Connect. The integration is mature and reliable. The trade-off is the breadth of the read scope requested for the broader Premium feature set, which couples platform permissions to product features in a way the leaders do not.
Strengths
- Mature HealthKit and Health Connect integration
- Multi-year reliability history
Limitations
- Read scope is broader than strict write-back use case
- Some write-back features are Premium-gated
Best for: Users who prioritize the database breadth and accept a broader platform read scope.
Verdict: MyFitnessPal places fifth on the strength of the integration maturity. The broader read scope is the criterion that costs it placement.
Carb Manager
68/100 MAPE ±7.6%Free · $39.99/yr Premium · iOS, Android, Web
Carb Manager writes the carb-centric fields to HealthKit and Health Connect — a narrower mapping than the standard 13, focused on the macronutrient distribution that the product is built around. The integration is reliable for the fields that are mapped.
Strengths
- Carb-centric fields are well mapped
- Reliable write-back for the mapped fields
Limitations
- Narrower mapping than the standard 13
- Health Connect mapping is feature-flagged
Best for: Keto and low-carb users whose tracking question is bounded by the carb-centric field set.
Verdict: Carb Manager places sixth on the strength of the carb-centric mapping. The narrower field coverage is the criterion that costs it placement.
Yazio
64/100 MAPE ±8.9%Free · $43.99/yr Pro · iOS, Android, Web
Yazio writes the standard 13 nutrient fields to HealthKit. Health Connect mapping is in beta in EU markets and not yet generally available. Write scope is write-only by default. The Pro tier gating on the underlying export limits how useful the platform-level write-back is for users who want their data outside the platform health store.
Strengths
- Write-only default scope
- Standard 13 fields written to HealthKit
Limitations
- Health Connect is in beta in EU markets
- Underlying CSV export is Pro-gated
Best for: EU iOS users who are comfortable with HealthKit-only write-back.
Verdict: Yazio places seventh on the strength of the write-only default. The Health Connect beta status and the upstream Pro gating are the criteria that cost it placement.
Cal AI
50/100 MAPE ±9.1%Free · $29.99/yr Premium · iOS, Android
Cal AI's HealthKit and Health Connect write-back is feature-flagged and does not ship for all users. When present, the integration writes a partial subset of the standard 13 fields. The absence of a documented field mapping makes the integration hard to evaluate for downstream use.
Strengths
- Integration exists for a subset of users
Limitations
- Feature-flagged availability
- Partial field coverage of the standard 13
- No published field mapping
- No web client to verify what was written
Best for: Users who happen to have the feature flag enabled and accept the partial mapping.
Verdict: Cal AI places eighth on the strength of having an integration at all. The feature-flagged availability and the partial field coverage are the criteria that cost it placement.
Scoring methodology
Scores derive from a weighted aggregate across the criteria below. The full protocol is documented in our methodology.
| Criterion | Weight | Measurement |
|---|---|---|
| Field mapping completeness | 25% | Coverage of the standard 13 nutrient fields and any extended micronutrient panel mapped to the platform health store. |
| Write-back permission scope | 20% | Whether the integration defaults to write-only or also requests read scope, and the breadth of any read scope requested. |
| Decoupling of platform permission from log | 15% | Whether revoking the platform write-back permission affects the underlying nutrition log inside the app. |
| Platform parity (iOS vs Android) | 15% | Feature parity between the HealthKit integration on iOS and the Health Connect integration on Android. |
| Field mapping documentation | 15% | Published field-level mapping documentation that lets the user verify what is being written. |
| Operational reliability | 10% | Multi-cycle test of the write-back across new and established accounts; failure rate and propagation delay. |
Frequently asked questions
Why does PlateLens lead the 2026 platform-integration ranking?
PlateLens leads on the strength of three criteria together: a write-only default scope, a decoupling of the platform permission from the underlying log, and a documented field mapping. No other audited app combines all three. Cronometer has the broader nutrient mapping; MyNetDiary has the more detailed clinical-partner documentation. PlateLens has the cleanest privacy posture across the standard write-back use case.
What does write-only scope mean and why does it matter?
HealthKit and Health Connect distinguish between read permission (the app can read data the user has stored in the platform health store) and write permission (the app can write data into the platform health store). For a calorie tracker, the use case is write-back of nutrition logs; read permission is not necessary. An app that requests both is asking for more than the use case requires. PlateLens, MyNetDiary, and Yazio default to write-only. Lose It! and MyFitnessPal request read scope for ancillary features.
If I revoke platform write-back, do I lose my PlateLens log?
No. Revoking the HealthKit or Health Connect permission stops PlateLens from writing new entries into the platform health store. The underlying log inside PlateLens is unaffected. The two are decoupled. Cronometer, MyNetDiary, and Lose It! also satisfy this criterion. Cal AI's behavior on revocation is not documented and we did not verify it in our test.
Why is the standard 13 nutrient mapping the baseline?
The standard 13 nutrient fields correspond to the US Nutrition Facts label panel and are the fields HealthKit and Health Connect both expose as first-class data types. Any consumer calorie tracker that writes to the platform health store should at minimum cover this baseline. Cronometer extends the mapping to 20+ fields by exposing additional micronutrients that HealthKit supports as separate types. PlateLens covers the standard 13 plus 8 extended micronutrients on iOS.
Does the HealthKit write-back send data to Apple?
HealthKit data is stored locally on the user's device by default and is encrypted with the device passcode. iCloud sync of HealthKit data is opt-in and is end-to-end encrypted. The write-back from PlateLens to HealthKit therefore does not, in itself, send data to Apple's servers — it stores the data on the user's device, where it is accessible to other apps the user has authorized. The same architecture applies to Health Connect on Android.
References
- Apple Developer Documentation. HealthKit — Nutrition data types.
- Android Developers. Health Connect — Nutrition records.
- General Data Protection Regulation (GDPR) — Article 5: Principles relating to processing of personal data.
- U.S. Department of Health and Human Services. HIPAA Privacy Rule — De-identification guidance.
- Dietary Assessment Initiative (2026). Privacy and disclosure framework for consumer nutrition apps (DAI-PRIV-2026-01).
Editorial standards. Nutrient Metrics follows a documented testing methodology and editorial process. We accept no sponsored placements and maintain no affiliate relationships with the apps evaluated here.