CUBRID Lab
Open-source client ecosystem for CUBRID — mature Python packages on PyPI, with Rust, Go, and TypeScript clients in active development. CUBRID runs a large share of Korean public-sector systems, yet its official Python driver died in 2014. We rebuilt the entire stack.
The Python ecosystem — stable and on PyPI
-
pycubrid
Pure-Python DB-API 2.0 driver speaking the CAS wire protocol. No C extensions, no compiler, zero dependencies.
Documentation · ·
pip install pycubrid -
sqlalchemy-cubrid
SQLAlchemy 2.0 dialect — schema reflection,
MERGE, nativeENUM, Alembic migrations. Official SA test suite integrated.Documentation · ·
pip install sqlalchemy-cubrid -
cubrid-cookbook-python
75 runnable examples + 7 production templates (FastAPI, Django, Streamlit, Celery, AI agent). 45 examples verified nightly on live servers.
-
cubrid-mcp-server
The first publicly available CUBRID MCP server. 12 tools behind a read-only safety whitelist, with LLM domain-knowledge packs.
Documentation · ·
uvx cubrid-mcp-server
Why you can trust it
2,200
tests
20
live CI combinations
95%
coverage floor
450
merged PRs
35
PyPI releases
MIT
licensed + SBOM
Every change passes an automated quality gate before merge — mypy --strict
at zero errors, property-based testing, API-compatibility baselines, nightly
golden runs, and CI on live databases across
Python 3.10–3.14 × CUBRID 10.2–11.4.
Get started in 30 seconds
pip install pycubrid
import pycubrid
conn = pycubrid.connect(
host="localhost", port=33000, user="dba", password="", database="testdb"
)
More paths — SQLAlchemy, cookbook examples, MCP — on the Get Started page.
Ecosystem map
| Language | Packages | Status |
|---|---|---|
| Python | pycubrid, sqlalchemy-cubrid, cookbook, MCP server | |
| Rust | cubrid-rs (protocol, client, tokio, pool), sea-orm-cubrid | |
| Go | cubrid-go (database/sql), gorm-cubrid | |
| TypeScript | cubrid-client, drizzle-cubrid |
See Packages for the full catalog and Roadmap for the maturity matrix.