Generator SDK

Framework Agnostic

Works with any frontend stack including plain HTML, React, Vue, and Svelte. Written in TypeScript with complete type definitions.

Ready Out of the Box

Ships in both ESM and UMD formats. Use it through npm or a CDN. All dependencies are bundled with no extra installation required.

Fully Featured

Includes authentication, cloud save and restore, history, credits, unified billing, export download, and other core platform capabilities.

Unified Billing

Built-in fallback logic from free quota to credits. The `withBilling` helper integrates billing in a single line of code.

Secure and Reliable

Automatic token recovery, automatic invalid-token handling, a unified `SdkError` system, and full data isolation across applications.

Vibe Coding Friendly

The CDN workflow is designed for AI coding scenarios, so a single prompt can generate a complete integration example.

SDK Landscape

Generator SDK is not just one package. The developer platform is composed of one runtime SDK for product features, one official shell layer, and one AI-facing toolchain for code generation and documentation retrieval.

What Each Piece Does

@atomm-developer/generator-sdk: the runtime SDK used by developers to integrate platform capabilities such as authentication, cloud save, history, credits, billing, and export.

generator-workbench: the official shell that hosts login, template actions, export entry, and other standard platform UI, so each generator does not need to rebuild the shell manually.

@atomm-developer/generator-sdk-mcp: the AI-facing MCP server that lets coding agents query documentation, inspect APIs, and generate runtime starters.

creating-generators skill: the reusable AI workflow package that teaches the agent how to ask for requirements, call the MCP tools, and generate standard generator projects.

Integration Map

There are two top-level integration modes:

Manual coding: integrate the runtime SDK yourself in a CDN, Vue, or React project.

AI coding: configure MCP plus skill, then let the AI generate or extend the project with the same platform capabilities.

The full decision flow lives in Integration Examples and the AI-specific step-by-step flow lives in AI Integration.

Start Here

If you are evaluating the platform as a whole, use this reading order:

Guide Overview — get the map of the guide set and the right reading path

Integration Decision Tree — choose the right integration path

Developer Journey — understand the end-to-end flow from local development to publish-ready integration

AI Vibe Coding Workflow — use MCP + skill to build generators through AI