For researchers & data scientists

Consented, de-identified, and actually usable.

A research workspace that pairs consented patient-journey data with live public oncology sources — queryable in a conversational assistant, explorable in a simulator, exportable in the formats your pipeline already speaks.

Live integrations

4 public sources

Oncology data bank

15 curated sources

Export formats

CSV · JSON · FHIR R4

Consent

Re-checked at download

What you get

Built around how you actually work

Research Assistant

A conversational assistant powered by Calanthe that queries live sources mid-answer — ClinicalTrials.gov, PubMed, cBioPortal, and NCI GDC — and shows you exactly what it queried.

Tumor-state simulator

Run the platform's rule engine on fully synthetic profiles — cancer type, stage, biomarkers, progression, recurrence — with zero patient data involved.

Cohort insights

De-identified aggregates across consented records: cancer-type and stage distributions, biomarker frequencies, treatment-response mix, mean tumor-state scores.

De-identified exports

Download CSV, JSON, or FHIR R4 bundles with pseudonymized ids that still group into cohorts. Consent is re-verified at download time and every export is audited.

Oncology data bank

15 curated public sources — registries, genomics portals, trial databases — ranked by cancer type, with evidence cards surfacing inside the analysis tools.

Consent-first by design

Patients opt in explicitly and can withdraw at any time. Names, emails, dates of birth, MRNs, and exact dates never enter the research plane.

From question to cohort

1

Explore public data

Ask the Research Assistant — it pulls live trials, literature, and genomics studies. No patient data required to start.

2

Simulate and aggregate

Test hypotheses against synthetic profiles, then check de-identified cohort distributions across consented records.

3

Export for your pipeline

Pull consented cohorts as CSV, JSON, or FHIR R4 — pseudonymized, audited, consent-checked.

De-identification you can defend

A research plane that never sees identity

Research access is architecturally separated from care access. The researcher role can never open an identified record — the API enforces it, and the audit log proves it.

  • Stripped before export: names, emails, DOB, MRN, exact dates, free-text identifiers
  • Preserved for science: diagnoses, biomarkers with values, treatment lines and responses, labs, symptom trajectories
  • Pseudonymized ids remain stable within an export so rows join into cohorts
  • Research consent is separate from care-team sharing and revocable independently
  • Every export creation and download is written to an append-only audit log
Research use only. Not medical advice. Public/derived data may be incomplete or outdated. Requires clinician/scientist review.

Put a live oncology workspace behind your next study

Explore the research tools on the demo environment, or contact us for researcher access to consented cohorts.

Not a researcher? Teloma for clinicians · Teloma for patients · Teloma for caregivers