DuckDB SQLAlchemy dialect

This project provides a production-ready SQLAlchemy dialect for DuckDB and MotherDuck. Use SQLAlchemy Core and ORM APIs with DuckDB locally or in MotherDuck without manual driver workarounds.

What you get

• SQLAlchemy Core, ORM, Alembic, and reflection support.
• MotherDuck helpers for attach modes, session hints, and read scaling.
• Sensible pooling defaults and optional performance tuning.
• Bulk insert helpers via Arrow or DataFrame registration.

Quick start

from sqlalchemy import Column, Integer, String, create_engine
from sqlalchemy.orm import declarative_base, Session

Base = declarative_base()

class User(Base):
    __tablename__ = "users"
    id = Column(Integer, primary_key=True)
    name = Column(String)

engine = create_engine("duckdb:///:memory:")
Base.metadata.create_all(engine)

with Session(engine) as session:
    session.add(User(name="Ada"))
    session.commit()

Next steps

• Connection strings: docs/connection-urls.md
• MotherDuck usage: docs/motherduck.md
• Configuration and pooling: docs/configuration.md
• OLAP helpers: docs/olap.md