Skip to content

koryki-java/core

Repository files navigation

Java ANTLR

Oracle PostgreSQL DuckDB

Verifiable AI-Assisted Semantic Querying for Relational Databases

The koryki.ai platform enables human-readable and supervised interaction with relational databases. It reduces complexity while preserving full control over what is queried and executed.

At its core is KQL (Koryki Query Language), a concise and human-readable language designed for ease of learning, interpretation, and validation. A well-defined grammar is key to making queries reliable and verifiable — for both humans and large language models.

From:

Find customers who have placed 
more than 10 orders in January 2023,
return companyname and count, sort by count.

To:

FIND customers c, orders o
FILTER count(o) > 10 AND 
    o.order_date BETWEEN 
      DATE '2023-01-01' AND DATE '2023-01-31'
FETCH c.company_name, count(o) DESC

The purpose of koryki is:

  • Shift control to human-centric queries
  • Simplify data analysis
  • Enhance workflows with AI while keeping full control

Read more, see sample query, or have a look at a guide to Koryki Query Language.

A demo application is available at: demo.koryki.ai.

KQL-Grammar Reference

Sub Projects

  • core: the koryki core library
  • duckdb: DuckDB dialect
  • northwind: Northwind sample database for testing purpose
  • oracle: Oracle dialect
  • postgresql: PostgreSQL dialect

Developer Documentation

Contribution

koryki is in early stage and open source under Apache 2.0 License

Any kind of feedback is welcome: info@koryki.ai

Packages

 
 
 

Contributors