sqlmermaid

Generate Mermaid ER diagrams from any SQLAlchemy-supported database — just pass a connection string.

erDiagram
  "complaints" {
     INTEGER about_dept FK
     INTEGER by_user FK
     DATETIME at_time
     VARCHAR-500 text "in index: comp_ind"
  }
  "departments" {
     INTEGER department_id PK
     VARCHAR-60 department_name
  }
  "user" {
     INTEGER user_id PK
     VARCHAR-16 user_name
     VARCHAR-60 email_address
     VARCHAR-50 nickname
  }
  "employees" {
     INTEGER employee_id PK
     VARCHAR-60 employee_name
     INTEGER employee_dept FK
  }
  "complaints" ||--|{ "departments" : "about_dept -> department_id"
  "complaints" ||--|{ "user" : "by_user -> user_id"
  "employees" ||--|{ "departments" : "employee_dept -> department_id"

Installation

pip install sqlmermaid

or uv tool install sqlmermaid

Usage

Command line

sqlmermaid <connection_string> [options]

# Print to stdout
sqlmermaid sqlite:///mydb.db

# Write Markdown
sqlmermaid sqlite:///mydb.db -o schema.md

# Open in browser
sqlmermaid postgresql://user:pass@localhost/mydb --browser

Python API

from sqlmermaid import get_mermaid, to_file, open_in_browser

# Mermaid diagram as a string
print(get_mermaid("sqlite:///mydb.db"))

# Write to a file
to_file("postgresql://user:pass@host/db", "schema.md")   # fenced Mermaid block
to_file("postgresql://user:pass@host/db", "schema.html") # standalone HTML page

# Open in browser
open_in_browser("sqlite:///mydb.db")

You can also pass a pre-populated sqlalchemy.MetaData object:

import sqlalchemy as sa
from sqlmermaid import get_mermaid

meta = sa.MetaData()
meta.reflect(bind=engine)
print(get_mermaid(meta))

Diagram features

• Columns annotated with PK and FK
• Index membership noted in column comments
• Foreign key relationships rendered as labelled edges
• Column size specifiers rewritten for Mermaid compatibility (VARCHAR(255) → VARCHAR-255, NUMERIC(10, 2) → NUMERIC-10-2)
• Views included when reflecting from a connection string

Output formats