What it does

This is the JavaScript SDK for Alan AI’s platform. You install it via npm, add a button to your web app, and users get a voice or text interface that can supposedly generate new features on the fly. The company calls this “Application-Level AI” — an embedded layer that reads your existing APIs, GUIs, and documentation, then writes code to match user requests at runtime.

The interesting bit

The pitch is “self-coding”: their proprietary “Three-Layer AI (3LAI)” architecture allegedly generates both business logic and UI in real time, with no developers needed. That’s a big claim, and the README treats it as established fact rather than aspiration. The SDK itself is essentially a thin client — the heavy lifting happens in Alan AI Studio, where you write dialog scripts in JavaScript and the platform handles the rest.

Key highlights

• Supports vanilla JS, React, Angular, Vue, Ember, and Electron
• Companion SDKs for iOS, Android, Flutter, Ionic, Cordova, React Native
• Includes working examples for React, Angular, Vue, Ember, and Electron
• Requires signing up for Alan AI Studio to build dialog scripts
• Distributed as @alan-ai/alan-sdk-web on npm

Caveats

• The “self-coding” and “no developers needed” claims are marketing language, not demonstrated in this repo
• The SDK repository contains no source code for the AI generation itself — it’s a client library for a proprietary cloud service
• Pricing, rate limits, and data handling policies aren’t mentioned in the README

Verdict

Worth a look if you’re building enterprise apps with heavy voice/text interaction needs and don’t mind vendor lock-in. Skip it if you need transparent, self-hosted AI or want to verify the “self-coding” claims independently.