Clinical research and medical educationThe institution's archive, ready to study and teach.
For researchers, faculty and innovation teams: counts, cohorts and de-identified case collections, with the source of every record and the institution's governance.

De-identified mode · ECG from 2021 to 2025
How many ECGs have a report mentioning atrial fibrillation or flutter?
3,412 ECGs, with no identifiable data.
Source Institutional ECG bank
Try it
From the question to the cohort and the classroom.
Two starting points, always de-identified.
Choose a question to see Kortex at work.
In the routine
Questions from research and teaching.
How many ECGs from 2021 to 2025 have a report mentioning atrial fibrillation or flutter?
Product coreBuild a class with pulmonary nodule cases from the last 5 years, without identifying anyone.
Product coreOf the patients with QTc above 470 ms, how many repeated the ECG within 30 days?
Product coreBuild an adult cohort with these criteria and show the coverage of each criterion.
Product directionExport the approved extract for the project, without identifiers.
Product directionExplore the cohort in Claude, with the same permissions as my profile.
Product direction
Your AI agentProduct direction
Explore the data in the agent you already use.
Connect Kortex to Claude, ChatGPT or another MCP-compatible client. The agent queries the data as a tool, with your profile's permissions and the de-identification it requires.
- The institution chooses which agents may connect
- Identifiable data never enters the model context
- Every query is recorded in the audit trail
Start with the archive the institution already has.
We talk about the sources, the formats and your research and teaching questions.