Omopforge

For OHDSI network-study sites and OMOP data engineers

Your OMOP ETL, already run once.

Omopforge maps source tables to OMOP CDM 5.4, compiles the ETL SQL for SQLite, PostgreSQL, SQL Server and DuckDB, runs it on a synthetic sample and scores the result with the official OHDSI data-quality thresholds. You review a checked draft instead of writing SQL from a blank file.

See the Synthea benchmark

111 / 111
person rows identical to the official ETL-Synthea output
98.4%
condition concept agreement with the codes stripped, label text only
252 of 256
DQD check thresholds passed on the output, failures listed by name

Public Synthea sample, run 2026-11-14, 9.5 s, 0 model calls. Every number, its method and the one-command re-run.

Built to be checked

Mappings with reasons

Each source column and code lands on a CDM field or standard concept with the evidence for the pick, and unmapped items are listed, not guessed.

SQL that has run

The ETL is generated for four dialects and executed on your sample before you see it; the row counts reconcile block by block.

A scorecard to show

Data-quality results arrive in DQD's own JSON shape, so they open in the tools your network already uses.

Run it on your own tables

Tell me when I can run this on my own tables

One email when it opens, nothing else. Privacy ยท What is stored