Datamimic is a deterministic synthetic data generation tool designed for CI/CD and analytics in regulated environments. The Community Edition (CE) is MIT-licensed, Python-native, and MCP-ready, enabling standalone use for synthetic data and PII-aware pseudonymization. It supports model-driven dataset generation without source data, with weighted state machines, composite references, and scriptable memstores.

Outputs include provenance logs and content hashes for audit. The Enterprise Platform adds governed workflows, PII scanning, role-based access, audit logging, scheduling, and multi-system execution. It supports databases like PostgreSQL, MySQL, Oracle, MS SQL, SQLite, MongoDB, and formats including CSV, JSON, XML, and XLSX.

Industry templates cover EDIFACT, SWIFT MT, HL7 v2.x, and FHIR. Deployments are available on-premise or air-gapped via podman-compose or Helm, with use in EU banking environments. The CLI allows agents to author, verify, and run data models using model.dm.json.

Install via pip install datamimic-ce and use .venv/bin/datamimic to avoid schema conflicts.