We’ll prove it before we claim it.
No accuracy figure appears on this site until VenTracer has been tested side by side with laboratory reference systems and the results are published, favourable or not.3 Here is the plan we will be held to.
In development · Pilot planned 2027 · Not for saleDesigned in India
The study, before the claims.
Example limits are shown to make the plan concrete; final limits will be fixed and registered in the protocol before data collection. How the methods and checks feed into it
| Element | Plan | Status |
|---|---|---|
| Before testing | ||
| Reference | A laboratory metabolic cart, tested on the same people | Planned |
| Bench | Metabolic simulator across 0.5–5 L/min V̇O₂. V̇E 20–100 L/min with the current flow element; 20–200 L/min with the high-flow element once built. Each element reported separately, including where it stops being accurate | Planned |
| Pre-registration | Protocol, acceptance limits and analysis plan registered publicly (for example on OSF) before the first participant is tested. Any deviation reported in the paper | Planned |
| Acceptance limits | Set before testing and justified from published reference-system agreement and the smallest worthwhile change in VO₂max, not chosen after seeing data (for example V̇O₂ bias within ±3% and 95% limits of agreement within ±7%; VT power within ±15 W) | Planned |
| Sample and population | Sample size justified in advance from the expected limits of agreement. Participants span VO₂max about 25–75 ml/kg/min, both sexes and ages 18–70 | Planned |
| Frozen versions | Firmware, algorithm and app versions frozen and named for the whole study; results apply only to those versions | Planned |
| During testing | ||
| Protocol | Seated rest, then an incremental ramp to exhaustion | Planned |
| Conditions | Cycling and treadmill running; a warm and humid room. V̇E above 150 L/min is tested only with the high-flow element, once it exists. Each flow element is validated and reported separately | Planned |
| Reference quality control | The reference cart is calibrated to the manufacturer’s procedure before each session and checked against a metabolic simulator or with Douglas bags. Its own error is reported | Planned |
| Measurement configuration | Where possible, both systems measure the same breaths (in series, with added dead space and resistance reported). Otherwise, repeated randomised-order tests. The configuration is stated in the paper | Planned |
| Reliability | Repeat tests on separate days, reported as typical error and smallest worthwhile change | Planned |
| After testing | ||
| Statistics | Bland–Altman bias and 95% limits of agreement (absolute and %), a test for proportional bias, and equivalence testing against the pre-registered limits. Correlation and ICC are reported but not used to judge agreement | Planned |
| Thresholds | Compared three ways: VenTracer algorithm vs cart (expert pick), expert picks on VenTracer data vs expert picks on cart data, and VenTracer algorithm vs experts on VenTracer data. Inter-rater agreement reported | Planned |
| Complete reporting | All tests reported, including failed tests, leak-flagged breaths, out-of-range breaths and dropouts. No post-hoc exclusions | Planned |
| Independence | Partner labs analyse their own data; the paper is submitted to a peer-reviewed journal whatever the result | Planned |
| Publication | Full results published, favourable or not, with de-identified data and analysis code, before any accuracy figure appears on this site | Committed |
None yet.
VenTracer has no published validation data. Until it does, readings from early units are experimental.1 When results exist, they will be listed here with the full paper, data and methods. The agreement chart for them is drawn on the methods page, and it stays empty until the data exist.
What VenTracer can’t do (yet)
- Claim accuracyNot until validation is published
- Cover maximal tests in trained athletesThe current flow element is characterised to about 85–100 L/min; a high-flow element is in development
- Find thresholds from short or broken testsA proper ramp is needed for a clear VT2
- Work through a leaking maskA seal check before every test and a leak check on every breath are planned
- Diagnose anythingIt is not a medical device4
From concept to published data.
- Now · 2026
Concept and architecture
System design, sensing plan, industrial design and brand.
- 2027 · planned
Engineering prototypes
First working analyzers on the bench and on athletes.
- 2027 · planned
Validation pilot
Units tested side by side with partner labs’ and clinics’ reference systems.
- After validation
Wider pilot
Units for coaches, athletes and clinics, once results are published.
Timelines are plans, not commitments, and may change with development and validation results.
Read the background.
How to evaluate a portable analyzer
Ten questions to ask every vendor, including us.
Read the guide → WearablesHow accurate is your watch?
What validation studies found against lab tests.
Read the guide → TestingVO₂max testing, explained
VO₂max vs VO₂peak, protocols and retesting.
Read the guide →Help us validate the first units.
Labs and longevity clinics that run a metabolic cart get priority as validation partners.
