Connectors
Start from the data you already have
Research data rarely lives in one place. It sits in a lakehouse, an EHR, claims files, an EDC, a CTMS and a dozen registry extracts. CuRE reads what those systems export, or the standards they speak, and maps what it reads into one governed OMOP record, so a new source is a connection to set up rather than a migration project
Data lakes and warehouses
- Spark SQL
- Amazon Redshift
CuRE is designed to run cohorts and analyses inside your own Databricks or Snowflake account, against OMOP tables you already hold, so person-level data stays where it is and only governed results come back. Cohort definitions already compile to Spark SQL for that purpose. On Amazon Redshift, CuRE's cohort SQL ran on a live warehouse, and 81 of 83 reference cohorts matched our Postgres reference exactly
OHDSI tools
- ATLAS
- HADES
- OHDSI WebAPI
Concept sets and cohort definitions are served through a WebAPI-compatible surface, so existing ATLAS and HADES tooling reads them with no migration
EHR and EMR
- FHIR R4
- SMART on FHIR
- Bulk FHIR export
- HL7v2
- C-CDA
Connect to EHRs through the standards they speak: SMART on FHIR launch and backend services, bulk FHIR export, HL7v2 feeds by file drop or MLLP listener, and C-CDA documents
Claims
- X12 837 claims
- X12 835 remittance
Payer claims and remittance files arrive as standard X12 interchanges and map to OMOP diagnoses, service lines and payer coverage periods
EDC
- Medidata Rave
- Veeva Vault EDC
Clinical data exported from Medidata Rave or Veeva Vault EDC is read, normalized and attached to its study, alongside data captured in CuRE itself
CTMS
- CTMS operational files
Any CTMS can send its operational data to CuRE as a file today. CuRE is also designed to connect directly to OnCore and Veeva CTMS and pull study operations into its own study records
Labs
- HL7v2 results
- FHIR Observations
- Scanned lab reports
Lab results arrive as HL7v2 messages or FHIR Observations. Scanned or PDF lab reports are read through a governed OCR and extraction pipeline before they land in OMOP
Registries and research datasets
- SAS
- Stata
- SPSS
- CSV and Excel
- CDISC SEND
Registry extracts and study datasets load as statistical-package or flat files, with an AI-assisted step that proposes how each column maps to OMOP
Patient-reported and device data
- Medidata Rave eCOA
- Veeva eCOA
eCOA exports are read the same way as EDC exports. For wearables and glucose monitors, CuRE is designed to connect to Validic, Fitbit and Dexcom and collect their device streams
Your own AI agents
- Model Context Protocol (MCP)
Agents you build can search the vocabulary and run and characterize cohorts through CuRE's MCP server, limited to the scopes they are granted and recorded in an audit trail
Data that flows into CuRE enters through one front door, so every source carries the same lineage and shows up in the same feed-health view. Don't see your system? Most platforms can export to one of these standards, and we will tell you which
Tell us where your data lives
Name the systems your team works with and we will map out how each one connects to CuRE