Explore

Try vasus.ai for yourself

Hands-on tools and demos. No signup, no data captured.

See vasus.ai in action — inside a research portal

Below is a fictitious research portal showing structured environmental exposure data for a 2,847-subject cardiovascular cohort. Pick any of the cohort's 4 sites to drill into 90-day trends and an Insight narrative for that location.

DEMO · fictitious portal · for illustration only
Powered by vasus.ai
Utrecht 2024 cohort (n=2,847)
Window: Jan–Mar 2026 Conditions: Cardiovascular, Respiratory Sites: 4 cities
Subject City Date PM₂.₅ EHSPI Linked citations
U-00184 Utrecht 03 May 14.2 6.2 PMID: 38274891PMID: 39102447
U-00185 Amsterdam 03 May 18.6 7 PMID: 38274891
U-00186 Utrecht 03 May 11.8 4.5 Lancet Planet. Health, 2024
U-00187 Rotterdam 03 May 22.1 7.8 PMID: 38274891PMID: 38995633
U-00188 The Hague 03 May 15.9 5.6 PMID: 39102447
⊕ Export CSV ↓ Equivalent API call showing 5 of 2,847 rows
Site detail
GET /v1/trends Utrecht · 90-day series
13.4 PM2.5 mean
22.1 PM2.5 p95
31.4 PM2.5 max
12 Days > WHO
Feb 1Mar 1Apr 1May 1
Linked PMIDs (raw signals)
PMID: 38274891PMID: 39102447PMID: 38995633
POST /v1/insight Utrecht · narrative
RISK: MODERATE
Synthesis
Cardiovascular sensitivity for the Utrecht site is moderately elevated, driven by repeated PM2.5 exceedance days clustered in late February. Pressure volatility events also tracked above baseline.
Key mechanism
Three consecutive days of PM2.5 above 25 µg/m³ correlates with a measurable rise in CV event rate in retrospective cohorts. The Utrecht site exceeded this threshold twice in the window.
Lancet Planet. Health, 2024 PMID: 38274891
Recommendation Cohort-level: include as a covariate in next-stage CV model. Site-level: flag late-Feb cluster for case-control comparison.
Both panels update together when you switch sites above.
Use case · Research institution

Cohort exposure & mechanism research

Pull longitudinal environmental exposure for any cohort site, with PMID-traceable signals — without manually merging WHO, ECMWF, and PubMed.

90+ days
Saved per cohort vs assembling exposure history manually from raw sources.
The problem

Researchers studying environment-disease relationships spend weeks merging air quality data, weather records, and PubMed-traceable mechanistic signals. The data engineering eats the science budget.

The solution

vasus.ai exposes GET /v1/trends for daily-resolution exposure series at any location and time window, plus POST /v1/insight for the synthesis layer with PMID-grounded mechanistic narrative. Python-friendly clients with batch endpoints for cohort-scale pulls.

The outcome

Researchers move from data-engineering to analysis in days, not weeks. Reproducible exposure inputs that anyone can re-pull. PMID-traceable mechanisms that survive peer review. Citable methodology section ready out-of-the-box.

Integration steps

Four steps to ship

1
Pull historical exposure
GET /v1/trends with cohort site lat/lon, variables list, window_days. Daily-resolution series.
2
Aggregate to your unit of analysis
Subject-day, subject-week, or whatever your study design needs. Response includes pre-computed aggregates.
3
Cite the mechanistic synthesis
POST /v1/insight — risk_level, summary, key_mechanisms with PMID-traceable citations for your methods section.
4
Document and reproduce
Citable methodology block referencing API version, weight vectors, and exact request parameters. Reviewers can re-pull.
Relevant sensitivities
cardiovascular respiratory migraine sleep allergies
Relevant for: epidemiology research groups, public health institutions, longitudinal cohort studies, environmental health PhDs.