ORMs & Query Builders

Drizzle ORM34.5k★

Your schema is defined in TypeScript, your queries look like SQL, and your IDE catches mistakes before your database does. Fully free under Apache 2.0. No paid tier, no cloud service. Drizzle Studio (a database GUI) is included. The team runs Drizzle Kit for migrations and schema management, also free. There's nothing to host. It's an npm package. Works with Postgres, MySQL, SQLite, Turso, Neon, PlanetScale, and more. Setup is `npm install drizzle-orm` plus a driver for your database. Migrations are handled by Drizzle Kit. Solo developers: if you know SQL and use TypeScript, Drizzle feels natural. The type inference is excellent. Change a column type and your app won't compile until you fix every query that touches it. Small to large teams: the SQL-like API means less abstraction to learn and fewer surprises. The catch: Drizzle is newer than Prisma and the ecosystem is smaller. Fewer tutorials, fewer examples, fewer community plugins. The relational query API (for nested/eager loading) is powerful but has a learning curve. Also, the documentation has gaps that the team is actively filling.

Sequelize30.3k★

Sequelize is the ORM for Node.js that works across Postgres, MySQL, MariaDB, SQLite, and SQL Server. Define your database tables as JavaScript models, and Sequelize handles the SQL: queries, joins, migrations, transactions, across Postgres, MySQL, MariaDB, SQLite, and SQL Server. It's the oldest and most feature-complete Node.js ORM. Associations, eager loading, raw queries when you need them, CLI for migrations and seeds. If you've used ActiveRecord (Rails) or Django's ORM, same concept. MIT. Fully free. The catch: Sequelize is showing its age. The API is verbose compared to newer ORMs like Drizzle or Prisma. TypeScript support was bolted on, not built in; the types work but feel like an afterthought. Drizzle ORM gives you type-safe queries with a much lighter API. Prisma gives you a schema-first approach with auto-generated types. Both are better choices for new projects in 2026. Sequelize is the right choice if you're maintaining an existing codebase that already uses it.

Kysely13.9k★

Kubernetes orchestrates containers across clusters of machines: scheduling, scaling, networking, storage, self-healing, all declaratively configured. It's not an ORM. It doesn't hide SQL behind methods. You write actual SQL, and TypeScript validates every piece of it. The difference from Prisma or Drizzle: Kysely doesn't generate SQL from a schema file or provide an abstraction layer. You write SQL directly using a builder pattern, and the type system ensures it's correct. If a column doesn't exist on a table, TypeScript errors. If you try to compare a string column to a number, TypeScript errors. It's SQL with a safety net. Completely free under MIT. No paid tier, no cloud service. It supports Postgres, MySQL, SQLite, and has community dialects for others. Drizzle ORM is the closest alternative, also type-safe, but with a more ORM-like API and built-in migrations. Prisma is easier to start with but generates heavier queries and has a larger runtime footprint. The catch: the learning curve is steeper than Prisma. You need to understand SQL and TypeScript generics. Schema changes require manual type updates (or a code generator). It doesn't handle migrations; you'll need a separate tool. And because it's closer to raw SQL, it's more code to write than Prisma's concise API.

SQLAlchemy11.9k★

SQLAlchemy is the Python toolkit for database access that works across Postgres, MySQL, SQLite, Oracle, and SQL Server. It works at two levels: a low-level 'Core' that maps Python expressions to SQL, and a high-level ORM that lets you work with database rows as Python objects. Fully free under MIT. This is foundational Python infrastructure; undersells its importance because virtually every Python web framework uses it. Flask, FastAPI, and dozens of others either integrate with or recommend SQLAlchemy. Version 2.0 modernized the API significantly with better typing and async support. The catch: the learning curve is real. SQLAlchemy gives you maximum control, which means maximum things to learn: sessions, engines, connection pools, lazy loading, eager loading, relationship cascades. If you just want a simple ORM without the complexity, Peewee or Tortoise ORM are simpler. And if your queries are complex enough, you'll end up writing raw SQL anyway through SQLAlchemy's text function. The abstraction helps most in the 80% of cases that are straightforward CRUD.