SEVN Vital
Make health information understandable across dimensions and time.
- Concept business
- SEVN VitalWellness-visualisation concept with synthetic personas: Ava (34), Malik (52), Ruth (67)
- First complete journey
- Explore a profile, change its time range, inspect its index and generate a grounded summary.
- Capabilities
- Website design, Visualisation, Conversational AI
- Status
- Interactive conceptSeed vital-v1 · resets daily 03:00 Sydney
Personal health information is scattered across apps, devices and appointments, each with its own units and gaps.
Single scores hide where a number came from; raw charts hide what matters.
SEVN Vital explores a navigable constellation of six dimensions over time, with a published index method, visible coverage and missing data, and plain-language summaries that cite their observations. Synthetic data only, and not a clinical score.
SEVN Vital
A consumer product website for SEVN Vital itself: interactive example, dimensions, method, data provenance, FAQs and a demo entry, clear about being synthetic and non-clinical.
- Home
- Interactive example
- Dimensions
- Method
- Data provenance
- FAQs
A SEVNworks concept website. SEVN Vital uses synthetic personas only.
SEVN Vital
Live preview loads when visible
Six steps,
each with a review.
Explore a profile, change its time range, inspect its index and generate a grounded summary. Every step names its input, what the system does, the rule or review that guards it, and what it hands on.
- 01 InputSynthetic profile ActionLoad 90 days of fictional observations Review or ruleUnits and dates explicit; gaps preserved OutputProfile snapshot
- 02 InputDimension scores ActionCalculate index from available dimensions Review or ruleMissing excluded; coverage falls visibly OutputIndex with coverage
- 03 InputTime range ActionCompare 7 / 30 / 90-day views Review or ruleChanges highlighted against source dates OutputComparison
- 04 InputHypothetical input ActionRun a labelled what-if scenario Review or ruleNever overwrites recorded history OutputScenario result
- 05 InputSnapshot and sources ActionDraft a plain-language summary Review or ruleEvery claim links to an observation OutputGrounded summary
- 06 InputApproved summary ActionExport snapshot, method and questions Review or ruleNo diagnosis or clinical claim OutputDownloadable synthetic summary
Workflow keys: observation_summary · snapshot_summary · adapter: native
What this build does,
and what it doesn’t.
Real in this build
- The index formula is deterministic and published: Σ(weight × available score) ÷ Σ(available weights).
- Coverage, missing dimensions and data age are shown separately: missing is never treated as zero.
- Weights and what-if inputs are editable and the maths updates live.
Simulated or saved
- All personas and observations are synthetic. No real medical records are accepted.
- Summaries are saved examples generated offline and grounded in the sample observations.
Limits
- Not a validated clinical score. No diagnosis, treatment, medication or risk prediction.
- Real-data use is a separate privacy, consent and clinical-definition scope.
Use it,
don’t just read it.
The demo workspace runs on fictional data with simulated external actions. No login. It resets to its sample baseline daily at 03:00 Sydney time.