Lyrebird · Realistic synthetic test databases, generated fully offline

Ark · Enterprise Test Data Management

Production-grade test databases.
Inside your own network boundary.

Ark classifies, masks, subsets, and provisions test databases inside your own private cloud or VPC. Engineering teams get realistic data in minutes; security and compliance teams keep absolute control of what leaves production.

  • Raw rows never leave your VPC
  • Outbound-only agent (mTLS)
  • PostgreSQL & MySQL
  • Self-hosted ephemeral targets
Ark web console: Business Objects with relational traversal rules

Platform Capabilities

The complete test data lifecycle.
Unified in one platform.

Ark handles discovery, masking, subsetting, and provisioning under one unified policy engine. Start with one module, or automate the entire lifecycle.

Discover & Assess
01 / CAPABILITY

Discover & Assess

Continuous PII detection across columns, unstructured text, and nested JSON. Risk scorecards, masking coverage metrics, and schema drift tracking.

  • Automated PII detection across structured & JSON columns
  • Schema drift monitoring before pipelines break
  • Risk scorecards and coverage metrics
Mask & Govern
02 / CAPABILITY

Mask & Govern

Deterministic, format-preserving masking policies executed locally beside the data. Governed Business Objects with relationship preservation.

  • Format-preserving hashing and replacement
  • Referential consistency across foreign-key tables
  • Versioned, auditable masking rules
Subset & Model
03 / CAPABILITY

Subset & Model

Relational subsetting with deterministic Anchors. Small production-shaped datasets that keep complete foreign-key dependency trees.

  • Deterministic Anchors for scenario-ready test cases
  • Deep relational traversal preserving parent-child integrity
  • Reduce multi-terabyte replicas to compact gigabyte slices
Provision & Automate
04 / CAPABILITY

Provision & Automate

Ephemeral PostgreSQL and MySQL test databases with automated TTL and ready DSN. Self-service from UI, CLI, SDK, or CI pipelines.

  • Short-lived sandbox environments on demand
  • Automatic TTL teardown to prevent cloud sprawl
  • Ready connection strings injected into test runners
Golden Snapshots
05 / CAPABILITY

Golden Snapshots

Versioned, reusable snapshots so environments are ready in minutes instead of waiting hours for slow database backup restores.

  • Instant environment initialization from snapshots
  • Deterministic state resets between test runs
  • Share certified baseline snapshots across teams
Compliance Reports
06 / CAPABILITY

Compliance Reports

Privacy compliance, risk scorecards, and residual review across KVKK, GDPR, CCPA, and CPRA aligned workflows. Export to PDF, XLSX, CSV, and JSON.

  • Residual audit inspects unmasked remnants
  • Framework-aligned compliance exports (PDF/CSV/JSON)
  • Immutable change history and job tracking

CLI & SDK Automation

Provision an ephemeral database,
run tests, and tear it down.

Engineers and CI/CD pipelines use Ark directly from the command line or SDKs to spin up a temporary, masked database, run integration tests against realistic data, and automatically destroy it upon completion.

  • ark-cli
  • Go SDK
  • JavaScript SDK
  • Python
  • Java
  • C# / .NET
  • ark-cli — pipeline automation

    # Provision an ephemeral test database from approved config

    $ ark-cli testenvs create --config "$ARK_CONFIG_ID" --ttl "45m" --wait

    [OK] Provisioning ephemeral PostgreSQL sandbox (ID: sandbox_7f2a)...

    [OK] Applying deterministic masking and customer subset...

    [READY] DSN: postgresql://ark_ci:***@10.42.0.18:5432/sandbox_7f2a


    # Execute your existing test runner against realistic data

    $ export DATABASE_URL="$ARK_TEST_DSN"

    $ npm run test:integration


    # Automatic cleanup after pipeline completes or TTL expires

    $ ark-cli testenvs destroy --id "sandbox_7f2a"

    [OK] Environment destroyed. Storage reclaimed.

    Zero-Trust Security Architecture

    Central control plane.
    In-network execution.

    Ark separates control from data. Ark Control plane manages policy, approvals, and audit trails. The Ark Agent runs entirely inside your VPC beside the data.

    • ✓ Outbound-only mTLS gRPC connection (no inbound firewall ports)
    • ✓ Read-only source access; raw rows never leave your network boundary
    • ✓ Immutable audit trail recording every job, access grant, and schema change
    SUBSETRA Ark Control

    Policy orchestration, drift monitoring, job schedules, audit trail

    Customer VPC / Private Network
    ARK In-VPC Agent

    Runs locally beside the data. Read-only connection to source DB.

    Raw Source DB → Masked Target DBs

    Built for Cross-Functional Teams

    Meet each team at its
    point of friction.

    Ark creates value immediately before a full enterprise rollout. Different teams enter from the same platform using their existing workflows.

    01

    Engineering & QA

    Realistic data on demand

    Provision relational, production-like databases through automated configs that are fast enough for CI and faithful enough for debugging.

    02

    Security & Privacy

    Stop PII before lower environments

    Classify sensitive fields, enforce deterministic masking, and keep reviews current as database changes introduce new risk.

    03

    Platform & DevOps

    Self-service data delivery

    Standardize how teams request prepared environments so test data stops depending on manual tickets, slow dumps, and stale configs.

    04

    Compliance & Legal

    Auditable privacy governance

    Automated KVKK, GDPR, and CCPA reporting, residual analysis, and audit trails proving that customer data never leaked into staging.

    Technical FAQ

    Common questions about Ark.

    Details on deployment, database compatibility, network connectivity, and compliance.

    How does the Ark Agent connect without inbound ports?

    The Ark Agent initiates an outbound connection over mTLS on port 443 to the Ark Control Plane. It functions purely as a worker polling for jobs, meaning no inbound ports or ingress routes need to be opened into your private network.

    Does raw production data ever leave our VPC?

    Never. The agent reads data from the source database, applies masking and relational subsetting algorithms in memory inside your VPC, and writes the sanitized output directly to local target test environments. Only job telemetry and metadata ever leave.

    Which database engines and platforms are supported?

    Ark supports PostgreSQL and MySQL. The agent runs as a Docker container or Linux binary on your own servers or in your private cloud.

    What compliance standards are supported?

    Ark provides automated mapping, continuous residual checks, and reporting aligned with KVKK, GDPR, CCPA, and CPRA. Audit reports can be generated in PDF, XLSX, CSV, and JSON.

    See Ark on your data

    Request a guided demo
    or the product brochure.

    Tell us about your database topology, compliance targets, and test environments. Brochure downloads: English (PDF) · Türkçe (PDF)

    Request a demo info@subsetra.com