SQL to SQLAlchemy Models

Paste CREATE TABLE SQL and get SQLAlchemy Python model classes instantly. Choose between SQLAlchemy 2.0\'s Mapped + mapped_column style (the modern default) or the classic Column approach. The converter maps SQL types to the right SQLAlchemy column types, applies primary_key, nullable, unique, and wires ForeignKey() from SQL REFERENCES clauses.

How to use the SQL to SQLAlchemy Models

Paste one or more CREATE TABLE statements (separated by semicolons) into the input box. Select the SQLAlchemy style you want — 2.0 Mapped uses Mapped[int] type annotations and mapped_column(), which is the recommended style for all new projects using SQLAlchemy 2.x. Classic Column uses the id = Column(Integer, primary_key=True) form compatible with SQLAlchemy 1.x and older tutorials.

Click Generate Models. The output is a Python module with a Base class and one model class per table. Each column is mapped to the appropriate SQLAlchemy type: INTEGER/SERIAL → Integer, VARCHAR(n)/CHAR(n) → String(n), TEXT → Text, BOOLEAN → Boolean, TIMESTAMP/DATETIME → DateTime, DECIMAL(p,s) → Numeric(p,s), JSONB/JSON → JSON, BYTEA/BLOB → LargeBinary.

FOREIGN KEY table constraints are picked up and rendered as ForeignKey("table.col") in the column definition. Use the Example button to load a two-table schema and see the output before pasting your own SQL.

What is SQLAlchemy?

SQLAlchemy is the most widely-used Python SQL toolkit and ORM. Its two main layers — the Core (expression language) and the ORM — give you fine-grained control over SQL while optionally providing a high-level object-mapping layer. Version 2.0, released in 2023, introduced the Mapped generic type annotation system which makes model definitions fully typed and introspectable by type checkers and IDE tools. If you are starting a new project or upgrading from 1.x, the 2.0 Mapped style is the recommended path.

In the 2.0 style a model column is declared as a class attribute with a Mapped[T] annotation and a mapped_column() assignment. The type annotation drives nullability — Mapped[int] is NOT NULL, Mapped[Optional[int]] is nullable. In the classic style the annotation is dropped and you write Column(Integer, ...) directly, which is compatible back to SQLAlchemy 1.0. Both styles work with the same declarative_base() / DeclarativeBase mechanism.

This converter handles the most common column types and constraints but does not emit relationship back-references (relationship()), column-level check constraints, or database-specific types like ARRAY or TSVECTOR. Those need to be added manually after conversion.

Common use cases

• Bootstrapping a new project — start from an existing SQL schema (from a DBA, a data warehouse, or a third-party tool) and get working Python model classes in seconds.
• Database reverse-engineering — dump CREATE TABLE from a legacy database and generate the ORM layer without writing boilerplate by hand.
• Learning SQLAlchemy — see how SQL column types and constraints map to SQLAlchemy constructs; compare the classic and 2.0 styles side-by-side.
• Multi-ORM projects — use the SQL DDL as the canonical schema definition (possibly generated from a Prisma schema) and generate models for different ORMs.
• Code generation pipelines — integrate the pattern into a CI step that keeps SQLAlchemy models in sync with the SQL schema checked into version control.