Technology

The stack underneath — languages, runtimes, infrastructure, principles.

Principles

What guides the stack choices

Boring is good

Mature, well-understood tools over fashionable ones. The platform processes identity data — novelty is a liability, not a feature.

Type-safety where it matters

Strong types in the core to catch identity-data bugs at compile time. Semantic types instead of generic strings throughout the model.

Open source where we can

Client libraries, internal frameworks (command4j, anythingworks) and several research tools live on GitHub.

Predictable scaling

Distribution where it's needed (Search Cluster), monoliths where it isn't — complexity is paid for, not chosen for fashion.

Long-lived APIs

Major versions stay stable for years — integration cost amortises over decades, not quarters.

Self-hostable

Every cloud product is shippable on-premise — same code, same algorithms, your infrastructure.

Stack

What runs underneath

Core engine

Java / JVM. Twenty years of name and identity-data algorithms — matchers, parsers, classifiers, the whole semantic-type model. Runs identically on cloud and on-premise.

Java / JVM

Search Cluster

Custom distributed index built for semantic-type search at PII scale — sub-50 ms search at 10 billion records, 100 billion transactions, 10k ops/sec. Multi-DC replication, no single point of failure.

Distributed index

REST & client surface

REST/JSON over HTTPS, fully described by an OpenAPI/Swagger specification. Typed client libraries for Java/JVM, Python and PHP, plus first-party Swagger UI for ad-hoc calls.

REST + OpenAPI

Frameworks & libraries

Internal frameworks command4j (command pattern for service APIs) and anythingworks (typed client foundations) — both open-sourced. The Java client builds on them directly.

Open source

Infrastructure

Cloud APIs run in EU data centres on dedicated infrastructure — not shared multi-tenant cloud functions. Off-site backup, cross-cluster replication, deterministic disaster-recovery procedures.

EU-based
Get in touch

Architecture or integration questions?

Happy to dig into how the platform fits your stack — networking, throughput planning, on-prem topology, anything.