SQLAlchemy to ER Diagram
Paste your SQLAlchemy declarative models and get an interactive ER diagram — tables from __tablename__, columns, and relations from ForeignKey.
At a glance
- What it is
- SQL to ER Diagram is a free, open-source browser tool that turns SQLAlchemy models into an interactive entity-relationship (ER) diagram.
- Input
- SQLAlchemy declarative model classes (Python source).
- Output
- An interactive ER diagram you can rearrange, then export as PNG, SVG, Mermaid, DBML or PlantUML, or share as a link.
- Cost
- Free. MIT-licensed open source. No account or signup.
- Privacy
- Runs entirely in your browser. The schema is never uploaded, and share links carry the diagram in the URL fragment, which is not sent to a server.
- Limits
- It never connects to your database itself: you paste schema text, or run one read-only catalog query in your own client and paste the result. There are no accounts, so there is no real-time multi-user editing; diagrams are kept in your browser, as a saved file, or as a share link. Layout stays smooth at around 300 tables.
How it works
- Copy your SQLAlchemy model classes (declarative
Base/db.Model). - Paste them into the editor — SQLAlchemy is auto-detected.
- Each model becomes a table;
ForeignKey("table.col")becomes a relation. - Arrange, hide tables, then export PNG/SVG or share a link.
Example
SQLAlchemy models like this:
class User(Base):
__tablename__ = "users"
id = Column(Integer, primary_key=True)
email = Column(String, unique=True)
class Post(Base):
__tablename__ = "posts"
id = Column(Integer, primary_key=True)
author_id = Column(Integer, ForeignKey("users.id"))
…renders two tables with a relation from posts.author_id → users.id. Try it with your own →
Guide
SQLAlchemy has two shapes of model code and both paste fine here: the classic Column(Integer, primary_key=True) declarative style, and the 2.0 typed style with Mapped[int] and mapped_column(...). Tables come from __tablename__ (or the class name if you omit it), columns from the Column/mapped_column calls, and every ForeignKey("users.id") — including composite ForeignKey("t.a"), ForeignKey("t.b") pairs — becomes a drawn relation.
What the parser deliberately ignores: relationship() back-references (the FK is what's drawn — relationship is Python-side only), hybrid properties, and methods. Nothing is executed; the models are read as text, so imports, custom base classes and decorators are all fine.
If your project uses Alembic, an even more exact route is alembic upgrade head --sql to print raw DDL and paste that instead — it captures server defaults and indexes the models don't show. But for most apps, pasting models.py directly is the two-second path to a diagram for your README or design doc.
Why use it
- Private: everything runs in your browser — nothing is uploaded or stored on a server.
- Interactive: drag tables, auto-arrange, hide noise, add notes, and manually link columns.
- Export a high-res PNG, vector SVG, or the schema as Mermaid / DBML / PlantUML code.
- Free & open source — no account, no sign-up.
FAQ
How do I visualize SQLAlchemy models?
Paste your declarative model classes. Tables come from __tablename__, columns from Column(...), and relations from ForeignKey.
Do I need to run my app?
No — it reads the model source directly in the browser, nothing is imported or executed.
Is my schema uploaded anywhere?
No. Everything runs locally in your browser — your schema is never uploaded to or stored on any server.
Is it free?
Yes. It is completely free and open source, with no account or sign-up required.