Reviews that took a month, not an afternoon
Every app below was bought at retail, worn or used for at least 31 days, benchmarked against reference hardware where the claim allows it, and torn down for trackers. Scores are out of 100 and traceable to six published axes.
Zenday App: the first tracker that knows what a dose is
The only app built around the real GLP-1 problem — protein on a suppressed appetite, and the week after a dose increase. One unpublished claim keeps it off the low nineties.
Sleepio: the most rigorously tested app in consumer health
26 clinical trials, 18 of them randomised, including a placebo-controlled RCT. The highest score we have given anything — and you probably can't just buy it.
Cronometer: the most accurate food database you can buy
USDA and NCCDB sourcing, staff-verified entries, 84 nutrients on the free tier. The more accurate product is also the cheaper one.
Headspace: the most-studied meditation app
Fourteen randomised trials against a competitor's one — half of them carrying a company conflict of interest.
Function Health: 114 biomarkers, and about six false alarms
The best data-per-dollar in consumer blood testing, attached to a results screen that doesn't tell you six flags were expected.
Calm: the best sleep content, the thinnest evidence
One published randomised trial, and the category's best sleep library. Know which of those you're buying.
MyFitnessPal: the world's food log, and its weakest link
Fastest logging in the category. Poor peer-reviewed validity across energy, carbs, protein, sugar, fibre and cholesterol.
Oura App 5.0: the quiet one still wins
Best-in-class sleep staging, a readiness score that finally explains itself, and a membership fee that stings a little less than it used to.
Whoop MG: brilliant coach, brutal contract
The strain model is the best in the business. The cancellation flow is the worst thing we tested all quarter.
Zoe: £300 of science, £300 of vibes
The test kit is real research. The app that follows is a food-scoring game that can't quite tell you why.
Everything we've scored
Live standings across the categories we track. Scores are re-tested every two major app releases or every 12 months, whichever comes first.
| App | Category | Score | Verdict | Last tested | 24-mo cost |
|---|---|---|---|---|---|
| Zenday App | GLP-1 companion | 89 | Our pick for GLP-1 users | Sep 2026 | $178 |
| Oura App 5.0 | Sleep | 84 | Strong buy for sleep nerds | Aug 2026 | $489 |
| Bearable | Symptom tracking | 80 | Best-in-class for chronic illness | Jul 2026 | $120 |
| Cronometer | Nutrition | 79 | The data nerd's food log | Sep 2026 | $0–120 |
| Headspace | Mind | 77 | The best-evidenced meditation app | Sep 2026 | $140 |
| Whoop 5.0 | Recovery | 76 | Buy with your eyes open | Aug 2026 | $478 |
| Function Health | Blood testing | 72 | Best value baseline panel | Sep 2026 | $730 |
| Levels | Metabolic | 71 | Good app, cancel after month three | Jul 2026 | $998 |
| AutoSleep | Sleep | 69 | Cheap, cheerful, noisy | May 2026 | $9 |
| Calm | Mind | 68 | Great sleep content, one published trial | Sep 2026 | $140 |
| Nutrisense | Metabolic | 66 | The dietitian is the product | Jul 2026 | $4,296 |
| InsideTracker | Blood testing | 65 | Better plan, much worse value | Sep 2026 | $1,276 |
| MyFitnessPal | Nutrition | 64 | Fast, social, and often wrong | Sep 2026 | $160 |
| Zoe | Metabolic | 61 | Wait for v4 | Aug 2026 | $1,240 |
| Sleepio | Sleep · CBT-I | 88 | Best-evidenced insomnia app anywhere | Sep 2026 | Often $0 via employer |
| Midi Health | Menopause | 83 | Best if you're insured | Sep 2026 | $0–$30/visit |
| Balance | Menopause | 81 | Best for the doctor's appointment | Sep 2026 | Free core |
| MacroFactor | Nutrition | 77 | Best adaptive macro coaching | Sep 2026 | $144 |
| Shotsy | GLP-1 dose | 76 | The best dose tracker, and only that | Sep 2026 | $80–100 |
| Noom GLP-1 Companion | GLP-1 programme | 74 | Free, and a funnel | Sep 2026 | $0 |
| MeAgain | GLP-1 companion | 72 | Flexible logging, no interpretation | Sep 2026 | ≈$519 |
| WeightWatchers GLP-1 | GLP-1 programme | 71 | A clinic membership with an app attached | Sep 2026 | ≈$1,750 |
| Phaze | GLP-1 companion | 70 | Local-first, feature-dense, great value | Sep 2026 | $70 |
| Quest Health | Blood testing | 76 | Real labs, no markup | Sep 2026 | ~$25+/test |
| Labcorp OnDemand | Blood testing | 75 | Same deal, different postcode | Sep 2026 | ~$25+/test |
| Evernow | Menopause | 75 | Async access, flat fee | Sep 2026 | $700–$1,180 |
| LetsGetChecked | Blood testing | 74 | Owns its labs | Sep 2026 | ~$50–200/test |
| Health & Her | Menopause | 74 | Best free self-help toolkit | Sep 2026 | Free |
| Sleep Cycle | Sleep | 73 | Fine without a wearable | Sep 2026 | $80 |
| Caria | Menopause | 72 | Structured guidance | Sep 2026 | Free tier |
| Everlywell | Blood testing | 70 | Broadest catalogue, split accountability | Sep 2026 | $49–299/kit |
| Superpower | Blood testing | 69 | Cheaper — until the price rise | Sep 2026 | $698 |
| Perry | Menopause | 68 | Best community | Sep 2026 | Free |
| Sleep Reset | Sleep · CBT-I | 67 | Accessible CBT-I, thin evidence | Sep 2026 | ~$150/programme |
| Lose It! | Nutrition | 66 | Friendly, cheap, wrong problem on a GLP-1 | Sep 2026 | $80 |
| Glapp | GLP-1 dose | 64 | Needs a second app to be useful | Sep 2026 | n/a |
| Cal AI | Nutrition | 52 | Pulled by Apple in April 2026 over billing | Sep 2026 | $60 |
| Rise | Sleep | 58 | One good idea, thin execution | Apr 2026 | $120 |
| Aura Mind | Mind | 44 | Avoid: five dark patterns, no export | Mar 2026 | $168 |
24-month cost includes required hardware, sensors and subscription at list price. No promo pricing, no annual-plan gymnastics.
In the lab
Eleven apps are mid-test. Day counts as of this morning.
| App | Testers | Day | Publishing |
|---|---|---|---|
| Eight Sleep Pod app | 2 | 24 / 31 | Sep 2026 |
| Function Health | 3 | 17 / 45 | Oct 2026 |
| MacroFactor | 2 | 9 / 31 | Oct 2026 |
| Apple Health (visionOS) | 4 | 3 / 31 | Nov 2026 |
Want us to test something? Readers set roughly a third of our schedule. Tell us what you're paying for and what you suspect, and it goes on the board.
Suggest an app85–100 · Best in category. We'd pay for it ourselves.
70–84 · Good, with named trade-offs.
55–69 · Works, but something meaningful is broken.
40–54 · Only for a specific person with a specific need.
Below 40 · We'd ask for a refund.