Personal signals, turned into experiments—not certainty.

Keep the interface from becoming a second science engine.

Health products become unreliable when views independently calculate, reinterpret or embellish underlying signals. BodySignal uses a canonical evidence provider so the same result, confidence and unknown state reach every presentation surface.

The interface explains evidence and next actions. It does not convert missing data into zero, imply causation from correlation or let generated copy overrule the deterministic result.

Make today useful without turning it into a dashboard wall.

Today prioritizes one clear reading and the next meaningful action. Patterns, history and experiments sit behind deliberate navigation and progressive disclosure, keeping the daily surface calm.

The sleep experience centers on a practical question: what actually changes your sleep? Prospective experiments preserve their schedule, context and outcome so a later verdict reflects what was planned—not a retroactive story.

• Uncluttered daily state
• Pattern exploration
• Prospective experiment schedule
• Typed unknown and insufficient-evidence states

Use fixtures for repeatability and label them honestly.

Synthetic QA scenarios make edge cases repeatable without exposing personal health records. They pass through the same canonical provider as real data, carry a visible simulated-data label and are blocked from share or export paths.

The canonical sleep-product scope completed 56 of 56 tests. Separately, the critical physical Apple Health comparison flow was exercised. Those are distinct evidence layers and are reported as such.

A health product whose restraint is part of the feature.

BodySignal connects native product design, a local evidence engine, deterministic experiments and device QA without asking the interface to sound more certain than the data.

That architecture creates room for future validation while protecting the user today—the strongest foundation a personal-evidence product can have.