npm version License: MIT TypeScript PostgreSQL
A type-safe ORM for PostgreSQL and TypeScript with automatic type inference and powerful query capabilities.
LINQ-Inspired Query Syntax: The query API is designed to feel familiar to developers coming from C# LINQ, with chainable methods like select(), where(), orderBy(), and groupBy(). You also get magic SQL string interpolation for when you need raw SQL power without sacrificing type safety.
PostgreSQL-First Philosophy: While other ORMs aim high and try to support all platforms, Linkgress is built exclusively for PostgreSQL. This allows it to leverage PostgreSQL's advanced features to the maximum—particularly in how collections and aggregations are retrieved using CTEs, LATERAL joins, JSON aggregations, and native PostgreSQL optimizations.
- Entity-First Approach - Define entities with
DbColumn<T>, no decorators needed - Fluent Configuration API - Intuitive
DbContextpattern with method chaining - Automatic Type Inference - Full TypeScript support without manual type annotations
- Nested Collection Queries - Query one-to-many relationships with CTE, LATERAL, or temp table strategies
- Type-Safe Aggregations -
count(),sum(),max(),min()return proper types - Powerful Filtering - Type-checked query conditions
- Prepared Statements - Build queries once, execute many times with named placeholders
- Fluent Update/Delete - Chain
.where().update()and.where().delete()with RETURNING support - Transaction Support - Safe, type-checked transactions
- Manual Migrations - File-based migrations with journal tracking, up/down support, and scaffolding
- Multiple Clients - Works with both
pgandpostgresnpm packages
npm install linkgress-orm postgres
See Installation Guide for detailed setup.
import { DbEntity, DbColumn } from 'linkgress-orm'; export class User extends DbEntity { id!: DbColumn<number>; username!: DbColumn<string>; email!: DbColumn<string>; posts?: Post[]; // Navigation property } export class Post extends DbEntity { id!: DbColumn<number>; title!: DbColumn<string>; userId!: DbColumn<number>; views!: DbColumn<number>; user?: User; // Navigation property }
import { DbContext, DbEntityTable, DbModelConfig, integer, varchar } from 'linkgress-orm'; export class AppDatabase extends DbContext { get users(): DbEntityTable<User> { return this.table(User); } get posts(): DbEntityTable<Post> { return this.table(Post); } protected override setupModel(model: DbModelConfig): void { model.entity(User, entity => { entity.toTable('users'); entity.property(e => e.id).hasType(integer('id').primaryKey().generatedAlwaysAsIdentity({ name: 'users_id_seq' })); entity.property(e => e.username).hasType(varchar('username', 100)).isRequired(); entity.property(e => e.email).hasType(varchar('email', 255)).isRequired(); entity.hasMany(e => e.posts, () => Post) .withForeignKey(p => p.userId) .withPrincipalKey(u => u.id); }); model.entity(Post, entity => { entity.toTable('posts'); entity.property(e => e.id).hasType(integer('id').primaryKey().generatedAlwaysAsIdentity({ name: 'posts_id_seq' })); entity.property(e => e.title).hasType(varchar('title', 200)).isRequired(); entity.property(e => e.userId).hasType(integer('user_id')).isRequired(); entity.property(e => e.views).hasType(integer('views')).hasDefaultValue(0); entity.hasOne(e => e.user, () => User) .withForeignKey(p => p.userId) .withPrincipalKey(u => u.id); }); } }
import { eq, gt } from 'linkgress-orm'; import { PostgresClient } from 'linkgress-orm'; // Create a database client with connection pooling const client = new PostgresClient('postgres://user:pass@localhost/db'); // Create a DbContext instance - reuse this across your application! const db = new AppDatabase(client); // Create schema await db.ensureCreated(); // Insert await db.users.insert({ username: 'alice', email: 'alice@example.com' }); // Query with filters const activeUsers = await db.users .where(u => eq(u.username, 'alice')) .toList(); // Nested collection query with aggregations const usersWithStats = await db.users .select(u => ({ username: u.username, postCount: u.posts.count(), // Automatic type inference - no casting! maxViews: u.posts.max(p => p.views), // Returns number | null posts: u.posts .select(p => ({ title: p.title, views: p.views })) .where(p => gt(p.views, 10)) .toList('posts'), })) .toList(); // Fluent update with RETURNING const updatedUsers = await db.users .where(u => eq(u.username, 'alice')) .update({ email: 'alice.new@example.com' }) .returning(u => ({ id: u.id, email: u.email })); // Fluent delete await db.users .where(u => eq(u.username, 'old_user')) .delete(); // Note: Only call dispose() when shutting down your application // For long-running apps (servers), keep the db instance alive // and dispose on process exit (see Connection Lifecycle docs)
Result is fully typed:
Array<{ username: string; postCount: number; maxViews: number | null; posts: Array<{ title: string; views: number }>; }>
- Getting Started Guide - Complete walkthrough for beginners
- Installation - Setup and installation instructions
- Database Clients - Choose between
pgandpostgres, connection pooling, and lifecycle management
- Schema Configuration - Entity configuration, relationships, and indexes
- Querying - Query data with type-safe filters, joins, aggregations, and more
- Insert/Update/Upsert/BULK - Insert, update, delete, and bulk operations
- Migrations - Automatic and manual migrations with journal tracking
- Collection Strategies - CTE, LATERAL, and temp table strategies for one-to-many queries
- CTEs (Common Table Expressions) - Build CTEs, query from a CTE root, and FULL OUTER / RIGHT / CROSS joins
- Subqueries - Using subqueries in your queries
- Custom Types - Create custom type mappers
- Node.js 16+
- TypeScript 5.0+
- PostgreSQL 12+
Contributions are welcome! Please see CONTRIBUTING.md for guidelines.
This project is licensed under the MIT License - see the LICENSE file for details.
- GitHub Issues - Report bugs or request features
- Discussions - Ask questions and share ideas
Crafted with ❤️ for developers who love type safety and clean APIs.