# DATAMIMIC

> Model-driven synthetic test data platform for regulated enterprises, generating deterministic, PII-safe datasets for banking, insurance, and compliance-heavy environments.

DATAMIMIC is a test data platform built by rapiddweller GmbH for banks, insurers, and other regulated enterprises. It generates deterministic, reproducible, PII-safe synthetic datasets on demand, integrating directly into CI/CD pipelines without production data ever leaving the customer's environment. The Community Edition (CE) is MIT-licensed and available on GitHub; the Enterprise Platform adds governed workflows, role-based access control, audit dashboards, and a Rust-accelerated execution engine.

## What It Is

DATAMIMIC is a model-driven test data generation and pseudonymization engine. Rather than scripting data by hand or copying production records, teams define abstract data models that capture business rules, referential integrity, and domain constraints. From those models, the platform produces byte-identical synthetic datasets on every run — same engine version, same model, same seed equals the same output, every machine, every time. The Community Edition covers single-system pipelines (PostgreSQL, MySQL, Oracle, MS SQL, SQLite, MongoDB, CSV, JSON, XML, XLSX), Python APIs, XML pipelines, a CLI, and an optional MCP server for AI agent integration. The Enterprise Platform extends this with multi-system coordinated execution, a Rust fastpath for billion-record workloads, ML/auto-regressive generators, a PII scanner, industry message templates (EDIFACT, SWIFT MT, HL7 v2.x, HL7 FHIR, ISO 20022), and on-premise or air-gapped deployment via podman-compose or Helm on OpenShift/Kubernetes.

## CE vs Enterprise Platform

The two editions share the DATAMIMIC DSL and determinism contract but use separate execution engines:

- **Community Edition (CE):** Python-native, MIT-licensed, multiprocessing with optional Ray distribution. Covers deterministic synthetic generation, manual-model pseudonymization, time-series generation, and MCP/IDE integration. Suitable for developers and data engineers in local, CI, or agent-driven workflows.
- **Enterprise Platform (EE):** Independently optimized execution engine with a Rust fastpath, ML/auto-regressive engine, keyset and manifest building from live DB schemas, and optimized distributed execution. Adds the full governance layer: RBAC, audit trails, DataWorkbench PII scanner, scheduled task runner, CI/CD integration (Jenkins, GitLab CI, Azure DevOps, Tricentis Tosca), and industry message template engine.

## Compliance and Audit Architecture

DATAMIMIC is designed around the traceability requirements of DORA, GDPR, BCBS 239, PCI DSS 4.0, and HIPAA. Every generation run is logged with a task ID, timestamps, model version, and a per-output content hash. Tasks are replayable from the seed, so any dataset can be reconstructed months later with byte-identical output. The platform supports two pseudonymization modes: seeded (deterministic, GDPR Art. 4(5) pseudonymization) for stable regression pipelines, and non-seeded (privacy-maximized, non-deterministic) for one-time data deliveries. The vendor notes that full anonymization status under GDPR depends on re-identification risk assessment by the data controller, not on individual field transformation alone.

## Domain Coverage and Format Support

CE ships with domain services across healthcare (Patient, Doctor, Hospital, MedicalDevice, MedicalProcedure), finance (Bank, BankAccount, CreditCard, Transaction), insurance, e-commerce, public sector, demographics, and common company data. All services are versioned and seeded. The Enterprise Platform adds industry-standard financial message generation:

- ISO 20022 (pacs.008, pain.001) and SWIFT MT (MT103, MT202 COV) from a single canonical payment model
- EDIFACT, HL7 v2.x, and HL7 FHIR via a spec-aware template engine
- Customers can upload their own spec files to extend the template catalogue between releases

## Update: DATAMIMIC CE 4.1.0

The latest Community Edition release is **4.1.0**, published on 2026-07-17. Version 4.0.0 introduced deterministic seeding for standalone literal generators (`<key generator="…">`), typed/pattern keys, `DateTimeGenerator`, and cross-page `unique` picks — making seed propagation machine-independent across the full DSL surface. The repository is actively maintained with CI-gated determinism architecture tests covering facade replay, service replay, and clock-drift detection. The MCP adapter (`pip install "datamimic-ce[mcp]"`) exposes canonical reference, scaffold, check, and bounded-run operations for AI agent integration, and the project lists MCP-ready status in its GitHub badges.

## Deployment and Integration Path

On-premise and air-gapped deployments are a first-class target. Container images are compact (server 250 MB, worker 750 MB, scheduler 150 MB) and require no internet connection at runtime — no telemetry, no license call-home, no cloud dependencies. The minimum system requirements are Linux with Docker/Podman or Kubernetes/OpenShift, 4+ CPU cores (>2 GHz), and 8 GB RAM. Teams integrate via a REST API and built-in scheduler, so test data arrives as part of the build from the first sprint. The vendor publishes case studies referencing a tier-1 European bank (Oracle, MongoDB, and Kafka pipelines), a global payments processor (real-time anonymization at millions of records per hour), and a public-sector education deployment (zero live student data across ~30 schemas) — all attributed as vendor-published claims.

## Features
- Deterministic, seed-based synthetic data generation
- Model-driven approach with reusable data models
- PII-safe pseudonymization (seeded and non-seeded modes)
- On-premise and air-gapped deployment via podman-compose or Helm
- ISO 20022 and SWIFT MT message generation
- EDIFACT, HL7 v2.x, and HL7 FHIR template engine (EE)
- CI/CD integration via REST API and built-in scheduler
- Role-based access control and audit dashboards (EE)
- PII scanner with probability-scored field detection (EE)
- Rust fastpath for high-volume generation (EE)
- ML/auto-regressive generators for complex distributions (EE)
- Multi-system coordinated execution: Oracle, MongoDB, Kafka (EE)
- Time-series generation with ISO 8601 intervals
- MCP server for AI agent integration
- Python API and XML pipeline support
- CLI with scaffold, lint, dry-run, and run commands
- Domain services: healthcare, finance, insurance, e-commerce, demographics
- Per-output content hash for audit re-execution
- GDPR, DORA, BCBS 239, PCI DSS 4.0, and HIPAA compliance support
- DataWorkbench visual field mapping and pseudonymization builder (EE)

## Integrations
PostgreSQL, MySQL, Oracle, MS SQL Server, SQLite, MongoDB, Apache Kafka, CSV, JSON, XML, XLSX, DbUnit XML, Fixed-width files, Jenkins, GitLab CI, Azure DevOps, Tricentis Tosca, Ray (distributed execution), Claude Code (MCP skill), Cursor (IDE)

## Platforms
LINUX, WEB, API, DEVELOPER_SDK, CLI

## Pricing
Open Source, Free tier available

## Version
4.1.0

## Links
- Website: https://datamimic.io
- Documentation: https://docs.datamimic.io
- Repository: https://github.com/rapiddweller/datamimic
- EveryDev.ai: https://www.everydev.ai/tools/datamimic