Skip to main content
Flock generates synthetic test data from your database schema: relationally consistent records that your integration engine and your database will both accept, as native HL7 v2, FHIR bundles, and FK-ordered SQL INSERTs. The data is synthetic from scratch; there is no production data to mask and no PHI at any point in the pipeline. Flock requires a Pidgeon account sign-in.

Before you start

  • Install Flock (or on macOS) and sign in.
  • Flock runs its engine in a local Bridge sidecar on localhost:5102. Schema introspection, learning, generation, and seeding all run locally.

One loop: schema in, seeded database out

1

Connect a schema

From the first-run screen, choose Connect a schema. Point Flock at a SQL DDL file or a live database connection. Flock reads tables, columns, and foreign-key relationships; that graph is what keeps the generated data FK-safe.
2

Configure the population

Open Population (under Generate). Pick a population size tier, set the age range and sex ratio, and optionally add edge-case profiles (Pediatric, Geriatric, Oncology, Maternal). The panel shows the generation limits for your plan before you run anything.
3

Generate

Run the generation. Records are produced in dependency order, so every foreign key resolves. Clinical values are grounded in public epidemiology sources (CDC, NHANES, Census), which is why the cohort’s condition mix looks like a real population instead of uniform noise.
4

Check the cohort

Open Analytics (under Learn) to see the distribution of the generated cohort, and Learn for the epidemiological references behind it.
5

Seed or export

Open Seed Database (under Generate). Export the population as FK-ordered SQL INSERTs, HL7 v2 messages, or FHIR bundles, or seed a target database directly.

When something fails

Next steps