We have written before about single-variable experiments — pick a creatine protocol, hold everything else constant, and watch a small set of outcomes for eight weeks. That structure works, and it is the right starting point for someone new to self-experimentation.
It is also unrealistic for how most lifter-experimenters actually live. By the time you have been at this for a year, you are simultaneously running a strength program, a sleep protocol, a magnesium-and-glycine evening routine, and a once-a-month CGM-tracked week. The variables stack. The dependent measurements stack. The system is multi-layered by necessity.
The question we get most often from readers is, what app handles all of this? The answer is: none of them. No single app handles all of it. The right framing is a composed stack — different tools for different layers — that together cover what you need without trying to be a single dashboard.
We use Prova ourselves (we make it) — and we use these other tools beside it. We disclose that involvement up-front. The composition below is the actual posture we run, not a synthetic recommendation list.
The composition principle
A health-experiment stack has five layers, and most apps try to handle two or three of them at most:
- Inputs — what you took, did, or changed. Supplements, training programs, sleep timing, food window.
- Outputs (performance) — what your body did under those inputs. Lifts, run paces, cognitive task scores.
- Outputs (subjective) — how you felt. Energy, sleep quality, mood, soreness, libido.
- Outputs (objective vitals) — what your devices captured automatically. HRV, RHR, sleep stages, VO2 max, body composition.
- Narrative — what you noticed, what you changed mid-experiment, what surprised you. The notes layer most apps neglect.
A reliable stack assigns each layer to the tool that handles it best, and accepts that the cost of composition is occasional cross-referencing between tools. The benefit is that no single tool sees all your data — which matters both for privacy posture and for analytical clarity.
Tool 1 — LiftProof, for the lifting layer (performance output)
When the dependent variable is can I lift more on this stack?, you need a lifting tracker that captures RPE, recovery, and long-term progression cleanly. We use LiftProof for this layer.
Three properties make it fit a multi-variable experiment posture:
On-device by default. Your lifting data is in Core Data on the phone, not on a vendor's backend. That keeps the lift log separate from your protocol log — which is the right separation for someone running structured experiments where co-correlation between supplement intake and lifting performance is the analytical question.
RPE first-class. Every set is RPE-tagged in the logger. RPE is the cheapest, fastest readout of recovery shift you have — and if your stack moves the needle on recovery, RPE before strength is where you will see it.
Watch + Live Activity. The Apple Watch app is full-featured, with five complications, four widgets, and Live Activity for active workouts. The lifting context lines up with HealthKit vitals without extra plumbing.
LiftProof has an honest essay on what v1.0 doesn't do, which we appreciate from a methodology-transparency standpoint — there is no cloud sync, no social, no Android at launch, no AI form-check. That clarity matters when you are composing a stack: you know exactly what the tool covers and what you have to handle elsewhere.
Tool 2 — Prova, for the protocol layer (inputs)
Prova is the supplement-and-protocol tracker we build. It handles the inputs layer — what you took, when, at what dose tier, how you felt about it. This is the layer most lifters under-track, partly because most lifting apps cannot accommodate the structure of a multi-week supplement protocol.
What goes in Prova: every compound in your stack with timing and dose tier, daily perceived effects, sleep and energy ratings, experiment notes when you change a variable. What does not go in Prova: lifts, vitals, narrative essays. Different tools, different jobs.
Pairing Prova with LiftProof gives you the two halves of the input/output equation — what you put in, what your body could do. Neither tool sees the other's data, which is a feature, not a bug, for self-experimenters who want their data posture deliberate.
Tool 3 — Bearable, for the symptom layer (subjective outputs)
Bearable handles the subjective-output layer — mood, energy, soreness, pain, libido, gut function, anxiety, sleep quality. The model is daily symptom check-ins with multi-scale ratings. It is cross-platform, has a serviceable correlation engine, and lets you correlate symptoms against factors you log.
We use Bearable for symptoms partly because Prova is intentionally narrower — Prova is for what you took, not for the full symptom inventory. Bearable's symptom-grain is much finer than what fits in Prova's UI, and the cross-correlation is the point: which symptoms moved, and when, against which inputs.
Bearable is cloud-default but encrypted at rest; the tradeoff is real but the data is fine-grained enough that we accept it for this layer. If you wanted a fully on-device symptom log, paper or Apple Notes with structure are the alternative.
Tool 4 — Apple Health, for the vitals layer (objective outputs)
Apple Health is the passive layer — HRV, RHR, sleep stages, VO2 max, body composition (if you have a scale that writes there), workouts captured automatically by Apple Watch. The point of this layer is that you do not interact with it manually. The vitals capture themselves; your job is to read them, not log them.
LiftProof, Prova, and Bearable can all be configured to write into or read from HealthKit, which keeps the vitals layer as the central correlation surface. Apple Health is also fully on-device — Apple does not move HealthKit data off the phone or to iCloud unencrypted. That makes it the right central hub for an on-device-preferring stack.
If you wear a Garmin or an Oura ring instead of an Apple Watch, the equivalent posture is to use their first-party app as the vitals layer and accept that some HealthKit-only integrations (like LiftProof's Watch-side Live Activity) will not be available.
Tool 5 — Notion or Apple Notes, for the narrative layer
This is the layer most experimenters skip and most regret skipping. The narrative layer is the why behind your data — what made you start the protocol, what you noticed mid-stream, what you changed and why, what surprised you, what you would do differently.
We use Notion for this — a single page per experiment, with the start date, the hypothesis, the variables held constant, weekly notes, mid-experiment changes, and the final read. Apple Notes works fine for the same job if Notion feels heavy.
The reason this layer matters: when you go to review a six-month protocol stack four months later, the data is meaningless without the context. Why did your training volume drop in week 4? Was it travel? Illness? A new variable you started? Without the narrative, you are guessing.
Composing it for yourself
The composition principle is simple: assign each layer to the tool that handles it best, accept that some manual cross-referencing is the cost, and gain the analytical clarity (and privacy posture) of not concentrating all your data in one place.
Start narrow if you are new. Pick two layers — inputs (Prova) and performance (LiftProof). Run that pairing for a quarter. Add the symptom layer (Bearable) once you have a feel for what you are tracking. Layer in vitals (Apple Health) and narrative (Notion) over the next quarter.
The lifters who get sustained value from this kind of stack are the ones who run it as a system, not a collection of apps. You are not "logging into" five tools — you are operating one composed instrument where each tool has a defined job.
Be the first to try Prova
We're building an app to track whether your health stack actually works. Join the waitlist.
Cross-link: the general-fitness perspective
Our sister site GetHealthyCalculators publishes lifter-focused listicles that ignore the multi-variable framing and focus on best-app-for-X comparisons. If you are not yet at the multi-variable experiment stage, those listicles are the more relevant entry point. For the peptide-and-lifting overlap audience, PeptideWise has a stack-composition variant that adds the regulatory-grey peptide layer to the same composition logic.
How LiftProof thinks about evaluating its own ranking
LiftProof publishes a meta-essay on its listicle methodology — how they audit themselves, where they refuse to rank #1, and what disclosure they apply. We linked to it because their transparency is what makes us comfortable featuring LiftProof as our preferred lifting layer here without it reading as a banner ad. Tools that publish their own methodology are easier to trust.
Disclosure: We make Prova. We feature LiftProof here because we use it ourselves; we evaluate fairly and disclose our involvement.
This article is for informational purposes only and does not constitute medical or training advice. Supplements and protocols may interact with medications or pre-existing conditions; consult a qualified healthcare professional before starting any new regimen.
Sources
- LiftProof — What's missing from LiftProof v1.0. liftproof.app, 2026. https://www.liftproof.app/blog/whats-missing-from-liftproof-v1/
- Apple Inc. Health app data privacy overview. Apple Support, 2026. https://support.apple.com/guide/iphone/iph44ac679fa/ios
- Helms, E. R., Cronin, J., Storey, A., & Zourdos, M. C. (2016). Application of the Repetitions in Reserve-Based Rating of Perceived Exertion Scale for Resistance Training. Strength & Conditioning Journal, 38(4), 42-49.