SuperMe
Open the team dashboard

The emotional OS that actually protects your privacy

Privacy is not a feature. It is the architecture.

SuperMe is designed to give teams the wellbeing insight they need — without collecting a single individual data point. Check-ins stay on the device; the server sees only differentially-private, k-anonymous team aggregates. There is nothing to breach, because by design it never exists.

Open the team dashboardReview privacy controls
  • No individual data — by design
  • Designed for GDPR & HIPAA alignment
  • Provable by design

Example aggregaten = 14 ≥ k = 8·ε 0.7

Insight you can use. Data you never hold.

Four layers designed to turn private, on-device check-ins into trustworthy team-level signal — stopping cold at the individual.

  1. On-device ML

    Personalized wellness plans are generated locally on each employee's device. Mood, stress, sleep, and journal entries never leave it.

    Designed so individual data never reaches a server

  2. Differential privacy

    Calibrated Laplace noise is added to every aggregate before it leaves the device — privacy designed to be provable, even against an unbounded attacker.

    ε-budgeted, composition-safe by design

  3. Federated aggregation

    Devices combine their signals through secure aggregation. The server receives one team number, assembled from many — never the parts.

    Suppressed below k = 8 contributors

  4. Team insights, zero individual data

    Leads see anonymized trends and risk patterns. “Team stress is up 20%” — never “John is struggling.” Patterns escalate, content never does.

    Human-confirmed, identity-free

  • 0

    Individual rows on the server

    By architecture, no individual table exists to breach or subpoena

  • k ≥ 8

    k-anonymity floor

    Designed to suppress — never estimate — aggregates below the floor

  • ε 0.7

    Differential-privacy budget

    Calibrated Laplace noise on every metric that leaves a device, by design

  • 100%

    Requests designed to be PII-scanned

    Privacy-guard middleware is built to reject individual data at the edge

Design guarantees of the SuperMe privacy architecture — an early-stage concept. These figures describe how the system is designed to behave, not audited production metrics.

A preview of the dashboard

Watch the trend. Never the person.

Every chart in SuperMe is an aggregate wrapped in a differential privacy band, gated by a k-anonymity floor. When a team is too small to protect, the data is suppressed — not estimated. Better no data than leaky data.

  • Differential-privacy confidence bands on every series
  • k-anonymity suppression below 8 contributors
  • Risk patterns escalate to a human — identity-free
Explore the demo dashboard

Team wellness index

7.4 / 10 +21% / 8 wks

n = 14 ≥ k = 8·ε 0.7 spent

Sample data (illustrative)

0369
Wk 1Wk 2Wk 3Wk 4Wk 5Wk 6Wk 7Wk 8
Sample data (illustrative) — how a 14-person team's aggregate would render. Shaded ribbon = differential-privacy noise band. Never an individual score.

Mental health is too important to be left to surveillance.

Trust shouldn't be claimed — it should be provable. SuperMe is designed so it can be. Step into the demo dashboard and see wellbeing the way it should be: aggregate, anonymous, and entirely yours.

Open the team dashboardSee monthly reports
SuperMe

SuperMe Wellness AI — an early-stage Acelera Intelligence Loonshot concept. Designed to be aggregated, differentially private, and k-anonymous only — never individual data. The dashboard is a demo with sample data.