SKAI Worldwide

AgensGraph

Relational stability, with graph flexibility

Request brochureContact sales

  • 6×

    Detection rate, Korea Customs

  • Grade 1

    GS Certification (Korea)

  • 20 billion

    Records analyzed at KB

  • Korea's First

    RDB + GDB hybrid

A Korean graph database
that reads relationships

A Korean hybrid graph database holding structured data and relationships in one engine. Run SQL and graph queries (Cypher) side by side.

The more complex the relationships,
the slower a relational DB runs

A relational database computes connections with JOINs.
The deeper the relationship, the slower it gets.
Relationships themselves are hard to work with.
A graph database stores relationships as they are, and traverses them.

Relationships stored as-is
Fast multi-hop traversal

Relationships are stored as they are, so speed holds
even across multi-hop traversal.

  • RDB + GDB hybrid

    Structured data and graphs in one engine.
    SQL and graph queries (Cypher) run together.

  • Large-scale graph processing

    Tens of billions of relationship records stored
    and traversed reliably.
    Performance holds on large graphs.

  • Apache AGE Top-Level Project (promoted 2022.06)

    Apache AGE, developed by SKAI and donated to the Apache Software Foundation, graduated to a Top-Level Project.

  • Officially shipped in Microsoft Azure

    Apache AGE, developed and donated by SKAI, now ships in Azure's managed PostgreSQL.

Problems that graphs solve

Where the relationship is the problem,
a graph database answers what a relational one cannot.

Fraud detection in finance

Network-based fraud detection — layered money flows,
mule accounts, voice phishing

Anomaly detection in the public sector

Fraudulent qualifications and improper claims
caught through relationship patterns

Knowledge structuring · evidence search

Scattered data woven into relationships and turned into knowledge — the foundation of Graph RAG

Network structure analysis

Complex networks across customers, organizations, and logistics —
traversed and visualized

The more complex the relationships, the more a graph database is worth

Works with your existing stack

  • 01Cypher
  • 02PostgreSQL
  • 03Standard SQL
  • 04JDBC · ODBC
  • 05Apache AGE
  • 06Heterogeneous DBs
  • 07BI tools
  • 08SQL + Cypher
  • 09On-Premise

Why AgensGraph

  • Hybrid

    Structured data and graphs in one engine —
    SQL and Cypher run together

  • Performance

    Tens of billions of relationship records processed reliably —
    fast multi-hop traversal

  • Standards

    RDB + GDB hybrid · Apache AGE Top-Level Project
    · commit rights held

  • Validation

    GS Certification Grade 1, Korea · 10 core patents ·
    Korea Customs detection · 20 billion records at KB

Built for regulated environments

Performance and international standards. The product already answers what a graph database review will ask.

  • Technology and standards

    SQL and Cypher run in a single engine

    RDB + GDB hybrid architecture

    Apache AGE named an Apache Top-Level Project

    Apache AGE shipped in Microsoft Azure

  • Performance and scale

    Tens of billions of graph records processed reliably

    Fast multi-hop relationship traversal

    Network anomaly detection and pattern analysis

    Redundancy and high-availability setup supported

  • Technical support · SLA

    Direct response from dedicated engineers at the Korean head office

    GS Certification Grade 1, Korea · 10 core patents

    Experience operating large-scale graphs

    Enterprise operating experience accumulated since 2013

Proven in production,
in live network analysis

Already running in detection and network-analysis systems across finance and the public sector.

Can a graph untangle the relationships in your data?

Contact sales

Frequently asked questions

Do we have to replace our relational database with a graph database?
No. AgensGraph runs SQL and graph queries (Cypher) in one engine. You keep your existing SQL assets and add graph where relationship analysis is needed.
What problems is a graph database for?
Any problem where relationships are the story: fraud and voice phishing, misconduct detection, knowledge graphs, and network analysis. It's in production at the Korea Customs Service, KB Kookmin Bank, and the National Tax Service.
How is it different from a foreign graph database?
A hybrid product that runs SQL and Cypher in a single engine. What sets it apart from foreign alternatives is Korean development, technical support, and ownership of the source code.
Does performance hold at scale?
KB Kookmin Bank runs a 20-billion-record graph on it. Multi-hop traversal stays fast as relationships go deeper.
Can it be used for knowledge graphs and RAG?
Yes. It connects scattered data through relationships and turns it into knowledge, forming the foundation for Graph RAG. It is the core engine of ONTOVIA, our knowledge AI product.
How do adoption and technical support work?
Dedicated engineers at the Korean head office respond directly, and PGTS maintenance backs deployment and operations with an SLA.

Evaluating a graph database?

Contact sales