Lattice SDKv3.2.1

Lattice SDK Documentation

Build type-safe, composable data layers for modern applications.

🔍Search documentation... (Ctrl+K)

Getting Started

Install Lattice SDK via your preferred package manager. The SDK requires Node.js 18 or later and supports both ESM and CommonJS module formats.

$ npm install @lattice/sdk

After installation, initialize your Lattice client by providing your schema definition and connection parameters. The client auto-discovers your data sources and builds an optimized query planner at startup.

// Initialize the client

import { LatticeClient } from '@lattice/sdk';

const client = new LatticeClient({

schema: './schema.lattice',

endpoint: 'https://api.example.com',

});

API Reference

client.query()

Execute a read operation against your data layer. Supports filtering, pagination, field selection, and nested relation loading with automatic batching.

Returns: Promise<QueryResult>

client.mutate()

Perform create, update, or delete operations with built-in optimistic locking. Mutations are validated against your schema before execution.

Returns: Promise<MutationResult>

client.subscribe()

Open a real-time subscription to data changes. Uses WebSocket transport with automatic reconnection and message ordering guarantees.

Returns: Subscription

client.batch()

Group multiple operations into an atomic batch. All operations succeed or fail together, with automatic rollback on partial failure.

Returns: Promise<BatchResult>

Code Examples

Fetch a paginated list of users with their associated organizations, applying a filter on account status.

const users = await client.query('User', {

filter: { status: 'active' },

include: ['organization'],

pagination: { page: 1, perPage: 25 },

orderBy: { createdAt: 'desc' },

});

// Result: { data: User[], meta: { total: 142, page: 1 } }

Changelog

v3.2.1 — February 2026

Fixed edge case in batch rollback when using nested transactions. Improved WebSocket reconnection stability.

v3.2.0 — January 2026

Added support for computed fields in schema definitions. New caching middleware with TTL-based invalidation.

v3.1.0 — December 2025

Introduced plugin system for extending client behavior. Added official Auth, Logger, and Metrics plugins.