Praenox reads your HRV, sleep architecture, and stress — then generates a therapeutic soundscape engineered for your exact physiological state. Not generic. Not approximate. Yours.
Two eras on one timeline. The left half is the raw Zepp baseline — tracked nights before Praenox. The right half is every night after Praenox started, brought right up to today. The dashed marker is the switch-over. No cherry-picking.
Every data point is entered as-is from the device — including the outliers you can see in the chart. The spike in deep sleep, the near-zero REM night, the recovery dip — all real, all included. If anything, the broken nights make the trend harder to achieve, not easier.
The gap in the baseline between 2022 and late 2025 is not missing data — the watch was worn but tracking was off and the app was never checked. No interest in the numbers at the time. That is exactly what unmanaged baseline looks like.
Free reacts to today. Pro learns your patterns and acts before decline. Developer adds the tools to test and inspect the engine. Aurum brings the complete adaptive Praenox experience to Android in a premium mobile edition.
— already downloaded
Requires a license key — buy on Buy Me a Coffee, key emailed within 24h
Not right for you? Email me — I'll respond within 48 hours.
— already downloaded
Requires a license key — buy on Buy Me a Coffee, key emailed within 24h
Not right for you? Email me — I'll respond within 48 hours.
Opens your email app with a message addressed to arlej.tech@gmail.com.
| Feature | Free | Pro | Developer | Aurum |
|---|---|---|---|---|
| Platform | Windows | Windows | Windows | Android |
| Biometric fingerprint synthesis | ✓ | ✓ | ✓ | ✓ |
| 6 biometric protocols | ✓ | ✓ | ✓ | ✓ |
| Audio export | MP3 | MP3 | MP3 | WAV |
| Personal baseline calibration | ✓ | ✓ | ✓ | ✓ |
| Progress charts & day comparison | ✓ | ✓ | ✓ | ✓ |
| Long-term pattern detection | ✗ | ✓ | ✓ | ✓ |
| Fusion engine (predictive) | ✗ | ✓ | ✓ | ✓ |
| 7-day recovery outlook | ✗ | ✗ | ✗ | ✓ |
| Context Check-in & History | ✗ | ✗ | ✗ | ✓ |
| Stability Mode | ✗ | ✓ | ✓ | ✓ |
| Sleep Debt Gate & Sleep Planner | ✗ | ✓ | ✓ | ✓ |
| HRV Velocity tracking | ✗ | ✓ | ✓ | ✓ |
| Streaks, shields & milestones | ✗ | ✓ | ✓ | ✓ |
| Outcome Learning / Learning Report | ✗ | ✓ | ✓ | ✓ |
| Multi-user profiles | ✗ | ✓ | ✓ | ✓ |
| Full profile ZIP portability | ✗ | ✗ | ✗ | ✓ |
| Background mobile playback & reminders | ✗ | ✗ | ✗ | ✓ |
| QA Runner (50+ automated tests) | ✗ | ✗ | ✓ | ✓ |
| Synthetic history simulator | ✗ | ✗ | ✓ | ✓ |
| Sound signature verification | ✗ | ✗ | ✓ | ✓ |
| Visual & runtime diagnostics | ✗ | ✗ | ✓ | ✓ |
| Development event log | ✗ | ✗ | ✓ | ✓ |
| Price | Free | $29 · one-time · lifetime | $59 · one-time · lifetime | $79.99 · one-time · lifetime |
Free reacts to each day in isolation. Pro learns your trajectory — it knows when your HRV is falling before you feel it, and forces the right protocol before decline arrives.
Why Free changes protocols daily, what Pro actually learns, how Fusion, Sleep Debt and streaks work, and what the Developer tools are actually for. Switch tabs to see each plan's answers.
The free version has no persistent Stability Mode and no Sleep Debt detection. It looks only at the health data you entered for that specific day — HRV, resting heart rate, deep sleep %, REM %, stress.
It then picks the protocol based on which single metric is furthest below your personal baseline (e.g. HRV too low → Parasympathetic Restore, deep sleep too low → Delta Induction, stress too high → Cortisol Reset).
Because your daily biometrics naturally fluctuate, the "worst metric" often changes day to day. The app is not broken — it is faithfully reflecting your daily variability.
Yes, but only your personal baselines, not behavioural patterns. It continuously updates your normal HRV, resting HR, typical deep sleep %, REM % and stress using all saved days, then uses them to calculate each new day's deficit.
What it does not do:
No. There is no spike-smoothing rule. If your HRV jumps from 50 ms to 80 ms overnight, the protocol for that day is based on the new value directly.
There is no "stable mode" that locks you into a single protocol for several days. If the protocol changes often, it's because your biometrics themselves are oscillating — the free version simply shows you that truth.
No, it's working as designed. The free version doesn't analyse chronic conditions — it reacts to each day's data independently.
High stress one day, low REM the next, good recovery the third produces Cortisol Reset → REM Restore → Maintenance. That's not a loop — it's a reflection of your changing state. To respond to underlying patterns instead of daily noise, that's what Pro was built for.
Six protocols, automatically selected from your biometric fingerprint:
The protocol label appears in the "Analysis" panel after generation.
Yes — try two things:
If you want the engine to learn and stop jumping, that's exactly what the paid version was designed for.
Not at all. The free version is excellent for understanding how your daily biometrics shape a therapeutic soundscape, experimenting with palettes and styles, and getting a fresh, data-driven session every night.
It just isn't designed to target chronic patterns or stabilise your protocol over time. For that, you'd need Pro.
No, it's working as designed. The app has learned your baselines and is now focusing on your most persistent issue. Priority order, highest to lowest:
If your sleep is regularly shorter than your baseline, the Sleep Debt Gate opens and forces "Sleep Debt Delta Recovery" until you catch up — three consecutive sufficient nights close the gate.
Fusion sits on top of the priority system. It can force a protocol in Stage 2 or 3 when an Act risk is detected, but it never bypasses Sleep Debt or Stability Mode.
Fusion prevents tomorrow's failure. Sleep Debt addresses today's deficit.
Stage 3 sub-labels by accuracy: < 55% → Calibrating, 55–69% → Tuning, ≥ 70% → Predictive.
Open Charts → "Your Optimal Sleep" card. It shows your personal baseline (e.g. 442 min) and confidence level (Low / Medium / High). This comes from your own history — not a fixed 7-hour rule.
They differ when your baselines improve over time. Both are correct — just different reference frames.
To protect the soundscape linked to that date. Locked fields become read-only and the delete button is disabled.
You can edit a locked day (confirmation required), but only future generations use the new values — the existing soundscape keeps its original snapshot forever.
Export awards points immediately and creates a pending confirmation in Milestones.
Preventive sessions (Fusion Stage 2+) use +13 listened, −8 skipped.
Shields — earn one shield every 7 streak days (max 2). If you miss 2+ days without opening the app, a shield automatically fires and your streak survives. You do nothing. The shield_user badge unlocks on first shield use.
The missing label at 40–55% is intentional — no more days are needed, you just need to keep marking LISTENED until accuracy climbs above 55%.
The other protocols (Cortisol Reset, REM Restore, Cardiac Calm, Maintenance) aren't gone — they're lower priority. To see them, sleep enough: 7 consecutive days with sleep ≥ your personal baseline (shown in Charts → Sleep Optimiser).
Once the Sleep Debt Gate closes, the engine cycles through the other protocols based on your daily fingerprint.
Every generated soundscape freezes a recovery snapshot at generation time. Once you mark a session LISTENED and the next day's entry is saved with a real HRV reading, Outcome Learning compares that next-day recovery against the snapshot and classifies the result:
Skipped sessions are excluded — a session you never heard says nothing about whether it worked. Once a protocol, style, palette category, layer, beat band, solfeggio frequency, duration, or story shape has at least 3 evaluated sessions, it earns a learned weight (0.6x–1.5x) that eases in gradually rather than snapping, so one rough night can't undo weeks of good data.
Fusion and Outcome Learning read the same saved history but track different things: Fusion predicts your next-day recovery and can force a protocol during a risk pattern; Outcome Learning tracks whether the soundscape itself actually helped. Fusion now also factors Outcome Learning's verdict on its forced protocol into its own confidence score. Both are shown side by side on the Learning Report tab, along with the active layer and precedence order for the selected day — Sleep Debt Delta Recovery first, then Fusion-forced protocol, then Outcome-learned weighting, then plain biometrics.
Nothing about the sound engine, Fusion, or scoring changes — Developer runs the exact same core as Pro. What's added is a dedicated Development tab with tooling to test and verify that core yourself: a QA Runner, a synthetic history simulator, sound signature verification, and visual/runtime diagnostics.
It's a different tier for a different kind of user, not a "more powerful" Pro.
An automated suite of 50+ logic tests covering context and generation, engagement, achievements, Fusion, data integrity, and outcome learning — plus a dedicated trust suite that checks things like snapshot freezing and per-user data isolation.
Run it on demand from the Development tab. It's the same suite used internally before every release.
Fusion Stage 2 needs 21+ HRV entries, Stage 3 needs 60+. Building that up with real nightly data takes weeks. The simulator generates a full synthetic biometric history so you can put the engine into any state — Fusion Stage 2/3, sleep debt, stability mode — instantly, without touching your real health data.
Every play and export captures a signature of the exact layers, seed, and parameters used. Verification confirms playback and export actually used the same generated soundscape — catching any drift between what was generated and what was heard or exported.
No. It's fully hidden and disabled in Pro — there's no button, no hidden shortcut, nothing to unlock later. It only exists in the Developer build.
No. If you just want the sound engine, Fusion, and the full recovery loop, Pro is the complete product. Developer exists for people who want to see how the machine works, verify it themselves, or run the same tests used internally — it doesn't change your day-to-day sessions at all.