NeuralNg Angular UI Expert MCP server

Documentation · Installation · Components · MCP server · Contributing

NeuralNg is an Angular UI component system designed first for code-generating agents and fully usable by human developers. It combines standalone, Signal-native APIs with strict types, accessible native semantics, SSR-safe behavior, headless styling and tree-shakable secondary entry points.

Beta: the current Core release is 0.1.0-beta.8. Public APIs may change

before the first stable release.

Why NeuralNg?

context, while the MCP server exposes component discovery and theme tools.

• Agent-readable: every public component ships focused README and llms.txt

Forms integration and deterministic hydration.

• Angular-native: standalone components, Signals, strict TypeScript, Angular

behavior and accessible names are part of component acceptance.

• Accessible by contract: native semantics, keyboard interaction, focus

tokens or complete unstyled control.

• Style without lock-in: use Neutral defaults, typed class slots, component

predictable, tree-shakable builds.

• Package what you use: import from dedicated secondary entry points for

Quick start

Install the component library and the framework-independent icon set:

npm install @neural-ng/core @neural-ng/icons

Load the icons and a bundled theme in your application styles:

@import '@neural-ng/icons/icons.css';
@import '@neural-ng/core/themes/neutral.css';

Import only the component entry points you need:

import { Component } from '@angular/core';
import { NeuralButton } from '@neural-ng/core/button';
import { NeuralInput } from '@neural-ng/core/input';

@Component({
  selector: 'app-agent-form',
  imports: [NeuralButton, NeuralInput],
  template: `
    <label for="agent-task">Task</label>
    <input neuralInput id="agent-task" placeholder="Describe the task" />
    <neural-button label="Run agent" icon="nt nt-sparkles" />
  `,
})
export class AgentForm {}

See the installation guide for application configuration, theming and SSR setup.

Packages

Shortened. The full README is on GitHub.