Govern tabular data
Import CSV/XLSX/TSV/TXT/Parquet or flat JSON/JSONL tables, inspect quality and preserve immutable versions, recipes and lineage.
AI tools for research · InsightAI Health Studio
Research with a record.
Explore your nonconfidential research tables, inspect methods and preserve the evidence behind each result.
DataCore is available in an authenticated research pilot for synthetic or nonconfidential user-owned tables. No personal, patient-identifying or confidential data. Research outputs are not clinical validation.
What DataCore can do
Inspect actual computation
Guided exploratory summaries, nine statistical tests, independent OLS/binomial/Poisson regression, classification/regression pipelines, PCA, clustering and anomaly exploration use actual CPU computation.
Review and export evidence
Grouped evaluation, baseline comparison, private artifacts, team roles and reviewed project-private research predictions.
The current server is qualified for bounded small research tables, one scientific job at a time. General large-data and institutional release gates remain open.
Read the DataCore guide →One platform, five phases
| Phase | Purpose | Current status |
|---|---|---|
| 1 · DataCore | Tabular research and shared core | Implemented; bounded nonconfidential research pilot |
| 2 · TextSense | Text research | Specification only; not installed |
| 3 · VisionSense | Image research | Specification only; not installed |
| 4 · AudioSense | Audio research | Specification only; not installed |
| 5 · AdaptFlow | Workflows and institutional integration | Specification only; not installed |
Later modules will reuse the same project permissions, versioned data and evidence contracts. No future-phase feature or API is enabled in this release.