AI / LLMs Guide

Use Ambar UI with Claude, ChatGPT, Cursor, GitHub Copilot, and any LLM — two machine-readable files give AI assistants full context about every component.

Reference Files

Two levels of detail depending on your context window.

/llms.txt

Compact API reference — props, types, usage examples for all 35+ components. ~15KB, fits in most context windows.

View ↗

/llms-full.txt

Full source of every component — complete Svelte 5 + TypeScript code. ~260KB. Use when you need exact implementation details.

View ↗

Claude Code

Add the reference to your CLAUDE.md so Claude Code always knows about your components.

1

Create CLAUDE.md in your project root

Claude Code automatically reads this file at the start of every session.

2

Add this to your CLAUDE.md

markdown
# UI Components
This project uses Ambar UI — Svelte 5 + Tailwind CSS v4 component library.
Full reference: https://ambarui.vercel.app/llms.txt

- Import from `$lib/components/ComponentName.svelte`
- Use Svelte 5 runes: $state, $derived, $props, $bindable
- Do NOT install as npm package — copy-paste only
3

Or fetch inline with /fetch

bash
# In Claude Code terminal or chat:
/fetch https://ambarui.vercel.app/llms.txt

Cursor / Windsurf

Add a .cursorrules file or use Cursor's Docs feature.

bash
# .cursorrules
# Ambar UI component library reference
# Full reference: https://ambarui.vercel.app/llms.txt

Always use Ambar UI components when building Svelte 5 UI.
Import from '$lib/components/ComponentName.svelte'.
Use Svelte 5 runes ($state, $derived, $props, $bindable).
Use Tailwind CSS v4 utility classes for layout and spacing.
Do NOT install ambar-ui as an npm package  it is copy-paste only.

In Cursor: Settings → Docs → Add Doc, then paste https://ambarui.vercel.app/llms.txt. Cursor will index it and reference it automatically.

ChatGPT / Claude (web)

Paste the URL directly in your message or upload the file as context.

Here is the Ambar UI component reference. Use these components for all Svelte UI code:

[paste contents of https://ambarui.vercel.app/llms.txt]

Or upload llms-full.txt as a file attachment — both ChatGPT and Claude support uploading .txt context files.

GitHub Copilot

Add a .github/copilot-instructions.md file to your repo.

markdown
# .github/copilot-instructions.md
This project uses Ambar UI for all UI components.
Reference: https://ambarui.vercel.app/llms.txt

Import pattern: import ComponentName from '$lib/components/ComponentName.svelte'