Migrate SQLAlchemy models to Prisma schema automatically

Moving a Python backend to a TypeScript or polyglot stack usually means Prisma. SQLAlchemy's declarative models are thorough — a decade of production tables covered in Column(), Index(), Enum(), and relationship() calls — but they describe structure in a format Prisma cannot read. Translating by hand is hours of work and almost always drops something.

SchemaForge converts SQLAlchemy model files to a Prisma schema as a reviewable starting point. Its shared representation preserves tables, columns, types, defaults, indexes, unique constraints, and enums; relationship details still need review before you apply generated output.

Install SchemaForge

# Install from source (PyPI publishing pending)
pip install git+https://github.com/Coding-Dev-Tools/schemaforge.git

A typical SQLAlchemy model

Suppose you have a models.py with a User table:

from sqlalchemy import Column, Integer, String, Boolean, DateTime, Enum, Index
from sqlalchemy.orm import DeclarativeBase
import enum

class Role(enum.Enum):
    admin = "admin"
    viewer = "viewer"
    editor = "editor"

class Base(DeclarativeBase):
    pass

class User(Base):
    __tablename__ = "users"
    __table_args__ = (
        Index("ix_users_email", "email", unique=True),
    )

    id        = Column(Integer, primary_key=True, autoincrement=True)
    email     = Column(String(255), nullable=False)
    name      = Column(String(120), nullable=True)
    role      = Column(Enum(Role), nullable=False, default=Role.viewer)
    is_active = Column(Boolean, nullable=False, server_default="true")
    created_at = Column(DateTime(timezone=True), server_default="now()")

Convert it

schemaforge convert --from sqlalchemy --to prisma --input models.py --output schema.prisma

For a directory of model files:

schemaforge convert --from sqlalchemy --to prisma --input ./models/ --output schema.prisma

What comes out

// schema.prisma — generated by SchemaForge

generator client {
  provider = "prisma-client-js"
}

datasource db {
  provider = "postgresql"
  url      = env("DATABASE_URL")
}

enum Role {
  admin
  viewer
  editor
}

model User {
  id         Int      @id @default(autoincrement())
  email      String   @db.VarChar(255)
  name       String?  @db.VarChar(120)
  role       Role     @default(viewer)
  is_active  Boolean  @default(true)
  created_at DateTime @default(now()) @db.Timestamptz

  @@index([email], map: "ix_users_email")
  @@unique([email])
}

Column names, nullable flags, string lengths, enum values, index names, and server defaults all carry over. The @@unique constraint is inferred from the unique=True on the Index.

What SchemaForge preserves

Batch conversion across a project

Most projects split models across multiple files. SchemaForge merges them into one output schema:

schemaforge convert \
  --from sqlalchemy \
  --to prisma \
  --input ./app/models/ \
  --output prisma/schema.prisma

Models from all Python files in the directory are combined. Duplicate table names cause a validation error so you spot them early.

Diff before you commit

After conversion, diff the generated schema against your existing Prisma file (if you have one) to see exactly what changed:

schemaforge diff \
  --format prisma \
  --a prisma/schema.prisma \
  --b prisma/schema.new.prisma

The diff output is field-level, not line-level — it surfaces added columns, changed types, and dropped constraints rather than whitespace noise.

Custom type mappings

SQLAlchemy custom types (TypeDecorator subclasses, JSONB, UUID, etc.) that SchemaForge does not recognise fall back to String by default. Override this with a YAML type-map:

# type-map.yaml
sqlalchemy_to_prisma:
  UUID: String   # or use @db.Uuid on Postgres
  JSONB: Json
  ARRAY: String  # Prisma has no native Array; Json is a common substitute

schemaforge convert \
  --from sqlalchemy \
  --to prisma \
  --input models.py \
  --output schema.prisma \
  --type-map type-map.yaml

VS Code extension

The SchemaForge VS Code extension gives you a live preview panel: open a SQLAlchemy model file, trigger "Convert to Prisma", and see the schema rendered side-by-side before you write it to disk.

Migration checklist

1. Run schemaforge convert on your model directory.
2. Review the output — verify that enum values, index names, and nullable flags match your expectations.
3. Add @relation blocks for any ForeignKey columns (see the warning above).
4. Run npx prisma validate to confirm the schema is syntactically correct.
5. Run npx prisma migrate diff between the generated schema and your current database to catch gaps before the first migration.

Other formats SchemaForge supports

If Prisma is not your target, SchemaForge supports 11 schema formats and 100 conversion directions: SQL DDL, Prisma, Drizzle, TypeORM, Django, SQLAlchemy, Alembic migrations, JSON Schema, GraphQL SDL, EF Core C#, and Scala case classes. Alembic is generator-only (an output target, not a source).

• Migrate TypeORM entities to Prisma
• Convert JSON Schema to Prisma
• SchemaForge overview — all 11 formats

Install SchemaForge →

SchemaForge is an MIT-licensed CLI tool built by Coding Dev Tools. Its shared representation preserves tables, columns, types, defaults, indexes, unique constraints, and enums. Review foreign-key constraints and ORM relationship fields after conversion.