The Standard Protocol for AI Software Generation.
Define your data in JSON/YAML. Run anywhere: Node.js, Browser, Edge. Empower LLMs to build hallucination-free backends.
ObjectQL is a universal data protocol and ORM engine designed for the AI Coding Era.
Traditional ORMs (TypeORM, Prisma) are built for humans writing code in IDEs. ObjectQL is built for LLMs generating logic. It uses rigid, declarative JSON/YAML Metadata and a strict JSON-AST Query Protocol to ensure AI Agents generate perfect, hallucination-free database operations.
Why ObjectQL?
- π€ AI-Native: Metadata-driven architecture provides the perfect context window for LLMs. No more parsing complex TypeScript classes or guessing SQL syntax.
- π‘οΈ Hallucination-Proof: The strict JSON protocol eliminates "imaginary methods" or invalid SQL syntax often generated by AI.
- π Universal Runtime: The core engine has zero dependencies on Node.js native modules. It runs in Browsers, Cloudflare Workers, and Deno.
- π Driver Agnostic: Switch between PostgreSQL, MongoDB, SQLite, or even a Remote API without changing your business logic.
ObjectQL is organized as a Monorepo to ensure modularity and universal compatibility.
| Package | Environment | Description |
|---|---|---|
@objectql/types |
Universal | The Contract. Pure TypeScript interfaces defining the protocol. |
@objectql/core |
Universal | The Engine. The runtime logic, validation, and repository pattern. |
@objectql/platform-node |
Node.js | Node.js platform utilities for file system integration, YAML loading, and plugin management. |
| Package | Environment | Description |
|---|---|---|
@objectql/driver-sql |
Node.js | SQL database driver (PostgreSQL, MySQL, SQLite, SQL Server) via Knex. |
@objectql/driver-mongo |
Node.js | MongoDB driver with native aggregation pipeline support. |
@objectql/driver-memory |
Universal | In-Memory Driver. Zero dependencies, perfect for testing and browser apps. |
@objectql/driver-localstorage |
Browser | Browser Storage. Persistent client-side storage using LocalStorage. |
@objectql/driver-fs |
Node.js | File system driver with JSON file-based persistent storage. |
@objectql/driver-excel |
Node.js | Excel file driver for using .xlsx spreadsheets as a data source. |
@objectql/driver-redis |
Node.js | Redis driver (example/template implementation for key-value stores). |
@objectql/sdk |
Universal | Remote HTTP Driver. Type-safe client for connecting to ObjectQL servers. |
| Package | Environment | Description |
|---|---|---|
@objectql/server |
Node.js | HTTP server adapter for Node.js, Express, Next.js with REST and metadata APIs. |
| Package | Environment | Description |
|---|---|---|
@objectql/cli |
Node.js | Command-line interface with AI-powered generation, dev server, and project management. |
@objectql/create |
Node.js | Project scaffolding tool (npm create @objectql@latest). |
vscode-objectql |
VS Code | Official VS Code extension with IntelliSense, validation, and snippets. |
The fastest way to start is using the CLI to scaffold a new project.
# Generate a new project
npm create @objectql@latest my-app
# Choose 'starter' for a complete example or 'hello-world' for minimal setupFor the best experience, install the official extension. It provides Intelligent IntelliSense and Real-time Validation for your YAML models.
- Search for "ObjectQL" in the VS Code Marketplace.
- Or open your new project, and VS Code will recommend it automatically via
.vscode/extensions.json.
ObjectQL uses YAML as the source of truth. Create a file src/objects/story.object.yml:
name: story
label: User Story
fields:
title:
type: text
required: true
status:
type: select
options: [draft, active, done]
default: draft
points:
type: number
scale: 0
default: 1Start the development server. ObjectQL will automatically load your YAML files and generate the database schema.
npm run devObjectQL isolates the "What" (Query) from the "How" (Execution).
- Supports PostgreSQL, MySQL, SQLite, SQL Server.
- Smart Hybrid Mode: Automatically maps defined fields to SQL columns and undefined fields to a
_jsonbcolumn. This gives you the speed of SQL with the flexibility of NoSQL.
- Native translation to Aggregation Pipelines.
- High performance for massive datasets.
- The Magic: You can run ObjectQL in the Browser.
- Instead of connecting to a DB, it connects to an ObjectQL Server API.
- The API usage remains exactly the same (
repo.find(...)), but it runs over HTTP.
- Zero dependencies - Pure JavaScript implementation
- Universal - Works in Node.js, Browser, Edge environments
- Perfect for testing, prototyping, and client-side state management
- See Browser Demo for live examples
- Browser-native persistence - Data survives page refreshes
- Built on Web Storage API
- Perfect for offline apps, PWAs, and user preferences
- See LocalStorage Demo for examples
ObjectQL runs natively in web browsers with zero backend required! This makes it perfect for:
- π Rapid Prototyping - Build UIs without server setup
- π± Offline-First Apps - PWAs with client-side data
- π Educational Tools - Interactive learning experiences
- π§ͺ Testing - Browser-based test environments
Try it now: Check out our interactive Browser Demo and LocalStorage Demo!
// Running ObjectQL in the browser - it's that simple!
import { ObjectQL } from '@objectql/core';
import { MemoryDriver } from '@objectql/driver-memory';
const driver = new MemoryDriver();
const app = new ObjectQL({ datasources: { default: driver } });
app.registerObject({
name: 'tasks',
fields: {
title: { type: 'text', required: true },
completed: { type: 'boolean', defaultValue: false }
}
});
await app.init();
// Use it just like on the server!
const ctx = app.createContext({ isSystem: true });
const tasks = ctx.object('tasks');
await tasks.create({ title: 'Build awesome app!' });ObjectQL's driver architecture supports custom database implementations. Potential databases include:
- Key-Value Stores: Redis, Memcached, etcd
- Search Engines: Elasticsearch, OpenSearch, Algolia
- Graph Databases: Neo4j, ArangoDB
- Time-Series: InfluxDB, TimescaleDB
- Cloud Databases: DynamoDB, Firestore, Cosmos DB
Want to add support for your database? Check out:
- Driver Extensibility Guide - Lists potential databases and their implementation complexity
- Implementing Custom Drivers - Step-by-step guide with code examples
- Redis Driver Example - Reference implementation
ObjectQL supports three distinct query interfaces, each optimized for different scenarios:
- JSON-DSL (Core): AI-friendly protocol, server-side logic, cross-driver compatibility
- REST API: Simple CRUD operations, mobile apps, third-party integrations
- GraphQL: Complex data graphs, modern SPAs, efficient multi-table fetching
Looking for best practices? Check out the Query Best Practices Guide for detailed comparisons, performance strategies, and optimization techniques.
Enhance your ObjectQL development experience with our official VSCode extension.
Features:
- π― Intelligent IntelliSense for
.object.yml,.validation.yml,.permission.yml,.app.yml - β Real-time JSON Schema validation with inline errors
- π 30+ code snippets for objects, fields, validations, hooks, and actions
- β‘ Quick commands to create new ObjectQL files from templates
- π¨ Custom file icons and syntax highlighting
- π Context-aware auto-completion
Installation:
cd packages/tools/vscode-objectql
npm install
npm run compile
npm run package
# Then install the generated .vsix file in VS CodeSee packages/tools/vscode-objectql/README.md for detailed documentation.
ObjectQL includes a powerful validation engine that runs universally.
# user.validation.yml
fields:
age:
min: 18
message: "Must be an adult"
# Cross-field validation supported!
end_date:
operator: ">"
compare_to: "start_date"
This validation logic runs:
- On the Client (Browser): For immediate UI feedback (via Object UI).
- On the Server: For data integrity security. Write once, validate everywhere.
Current Version: 3.0.0
Overall Completion: ~75%
- Roadmap - Long-term vision, milestones, and strategic direction
- Development Plan - Detailed 6-month development plan
- Project Status - Current state and metrics
- Contributing Guide - How to contribute
- β Core Protocol & Runtime: 90%
- β Data Drivers (SQL/Mongo/Memory/LocalStorage): 85%
- β Developer Tools (CLI, VSCode Extension): 85%
- π Workflow Engine: 35%
- π Security & Permissions: 70%
- π Documentation: 75%
If you fork or clone the repository to contribute or run examples from source:
-
Setup Workspace
git clone https://github.com/objectql/objectql.git cd objectql npm install -g pnpm pnpm install -
Build Packages You must build the core libraries before running examples, as they rely on local workspace builds.
pnpm build
-
Run Examples
These examples run as scripts to demonstrate the ObjectQL Core Engine capabilities (Validation, CRUD, Logic Hooks). They use an in-memory SQLite database.
Starter (Project Tracker):
# Starts the API Server (http://localhost:3000) pnpm --filter @objectql/example-project-tracker start # Note: Sample data (projects.data.yml) is automatically loaded on startup
Enterprise ERP:
pnpm --filter @objectql/example-enterprise-erp start # Output: Plugin initialization, Employee creation logs, Audit trails
ObjectQL is open-source software licensed under the MIT License.
You can use it freely in personal, commercial, or open-source projects.
ObjectQL (Data) β’ ObjectOS (System) β’ Object UI (View)