drizzle-docs-generator

What drizzle-docs-generator is

drizzle-docs-generator is a free command-line tool for producing database documentation from Drizzle ORM schemas. It generates Markdown by default or DBML, giving Drizzle users a way to turn schema definitions into a data dictionary. Markdown can be produced as one file or as multiple files, and can include a table list, a Mermaid ER diagram, and separate documentation for each table. An option can exclude the diagram. DBML output can be imported into dbdiagram.io to view ER diagrams and table definitions.

The tool extracts JSDoc comments from table and column definitions, including them in generated documentation and converting them into DBML notes. It imports schema files from a directory and supports extensionless imports. It detects relations declared with `defineRelations()` or `relations()`, and can fall back to foreign key constraints. A watch option regenerates documentation when schema files change, which the maker describes using in CI/CD workflows. Documented dialects are PostgreSQL, MySQL, and SQLite. The README recommends local installation as a development dependency and also documents global installation with npm or pnpm. The project is licensed under MIT and intended for self-hosted deployment.

Who it suits

This tool suits Drizzle ORM users who want documentation generated from their schema definitions and kept current as those files change. It is relevant to teams maintaining relational database schemas in PostgreSQL, MySQL, or SQLite and preferring a command-line workflow.

Getting started

1. Open the project on GitHub.
2. Install it locally as a development dependency, or use the documented global installation route with npm or pnpm.
3. Import schema files from a directory.
4. Choose Markdown or DBML output and configure Markdown file and diagram options if needed.
5. Enable the watch option to regenerate documentation when schema files change.

Verdict

Pick drizzle-docs-generator if you use Drizzle ORM and want schema-derived Markdown or DBML, including extracted comments and relation details. Look elsewhere if your database dialect is outside PostgreSQL, MySQL, and SQLite or you do not want a self-hosted CLI workflow.