dbcube

The fastest way to work with MySQL, PostgreSQL, SQLite and MongoDB in Node.js.

DBCube runs your queries through a persistent Rust engine (daemon mode) with warm connection pools — sub-millisecond query dispatch instead of paying process-spawn overhead, plus a fluent query builder, real transactions and schema-as-code with .cube files.

Installation

npm install dbcube
npx dbcube init   # scaffolds dbcube.config.js + dbcube/ + example schema

Configuration

dbcube.config.js in your project root:

module.exports = function (config) {
    config.set({
        databases: {
            myapp: {
                type: "mysql", // mysql | postgres | sqlite | mongodb
                config: {
                    HOST: process.env.DB_HOST,
                    USER: process.env.DB_USER,
                    PASSWORD: process.env.DB_PASSWORD,
                    DATABASE: process.env.DB_NAME,
                    PORT: 3306
                }
            }
        }
    });
};

Usage

const dbcube = require('dbcube');
const db = dbcube.database('myapp');

// Queries
const users = await db.table('users')
    .where('status', '=', 'active')
    .orderBy('created_at', 'DESC')
    .limit(10)
    .get();

// Writes (update/delete always require a WHERE)
await db.table('users').insert([{ name: 'Ada', email: '[email protected]' }]);
await db.table('users').where('id', '=', 1).update({ status: 'inactive' });

// Transactions — everything commits or everything rolls back
await db.transaction(async (trx) => {
    await trx.table('accounts').where('id', '=', 1).update({ balance: 50 });
    await trx.table('accounts').where('id', '=', 2).update({ balance: 150 });
});

// Raw SQL escape hatch (bound parameters)
const rows = await db.raw('SELECT * FROM users WHERE age > ?', [25]);

// Eager loading from your .cube foreign keys — one batched query, no N+1
const usersWithOrders = await db.table('users').with('orders').get();

// Pagination, upsert, atomic counters…
const page = await db.table('posts').paginate(2, 20);
await db.table('settings').upsert([{ key: 'theme', value: 'dark' }], ['key']);
await db.table('products').where('id', '=', 7).increment('stock', 5);

Advanced configuration

Tune the connection pool and daemon per database:

myapp: {
    type: "mysql",
    config: { /* ... */ },
    pool: {
        maxConnections: 20,      // default 5 (sqlite: 10)
        minConnections: 4,       // default 2
        acquireTimeoutMs: 5000,  // default 3000
        idleTimeoutMs: 600000    // default 3600000
    },
    daemon: {
        enabled: true,           // false → one-shot mode (transactions need the daemon)
        requestTimeoutMs: 60000  // default 30000
    }
}

TypeScript

Generate interfaces from your schema and get fully typed queries:

// npx dbcube generate → dbcube/types.ts
import type { User } from './dbcube/types';

const users = await db.table<User>('users').where('age', '>', 25).get(); // User[]
const one   = await db.table<User>('users').first();                     // User | null

Highlights

• Daemon mode: persistent TCP engine, ~0.3 ms per query (disable with DBCUBE_DAEMON=0)
• Transactions with automatic rollback on error — MySQL, PostgreSQL, SQLite and MongoDB (replica set)
• Schema as code: define tables in .table.cube files, apply with npx dbcube run table:refresh
• Migrations with history and rollback (.alter.cube + migrate:status / migrate:rollback)
• TypeScript types generated from your schema: npx dbcube generate
• Aggregations (count/sum/avg/max/min), exists(), chunk(), whereIn/whereNotIn, having, selectRaw, computed fields and runtime triggers
• Runtime triggers scale horizontally: they run in JS in the process that performs the write, so with N service replicas each operation fires its trigger exactly once — no distributed coordination needed

Documentation

Full docs: https://dbcube.org — CLI reference: npx dbcube help

License

MIT