Documentation

AI Development with SVAR

Build applications with SVAR UI components using AI coding assistants such as Claude Code, Cursor, Codex CLI, Antigravity CLI, GitHub Copilot, and ChatGPT.

ai

In the modern development landscape, AI automation tools have become essential for efficient coding and
problem-solving. SVAR provides three ways to help AI agents generate accurate code:

MCP Server

Direct access to the latest SVAR documentation and examples

Skills

Agent-native instructions for supported coding tools

Context Packs

Markdown files containing condensed SVAR documentation

MCP Server

We offer a Model Context Protocol (MCP) server, which gives AI assistants access to the latest SVAR documentation,
API references, and code examples. This helps AI tools generate more accurate code and reduces the risk of hallucinations.

Server URL

Why use the SVAR MCP Server?

SVAR MCP server makes generated code more accurate and reduces the risk of using outdated APIs. It provides:

  • Access to the latest documentation
  • Current API references
  • Up-to-date code examples
  • Component-specific implementation guidance
  • Retrieval-based answers generated from official SVAR resources

For installation instructions and supported integrations, see the MCP Server documentation

Supported tools: Claude Code, Cursor, OpenAI Codex CLI, Antigravity CLI, GitHub Copilot, ChatGPT or any other MCP-compatible AI environment.

mcp mcp

Skills

SVAR skills work with AI coding agents that support the SKILL.md convention, including Claude Code,
Cursor, Codex CLI, and Antigravity CLI. Run the following command to install the skill into your project,
or visit the skill repository and copy files from there.

React logo

React


npx skills add svar-widgets/skills --skill svar-react
Svelte logo

Svelte


npx skills add svar-widgets/skills --skill svar-svelte
Vue logo

Vue


npx skills add svar-widgets/skills --skill svar-vue

Skills are loaded automatically when the AI agent decides they match the task. To explicitly activate a skill:

In Claude Code

start your prompt with /svar-svelte

In Codex CLI

start your prompt with $svar-svelte

Supported tools: Claude Code, Cursor, OpenAI Codex CLI, Antigravity CLI

Context Packs

Context Packs are markdown files with the essential information needed to work with specific SVAR components. They are portable documentation bundles for AI tools that do not support Skills or MCP integrations. Find context packs in the guides folder and attach them to prompts, upload to AI assistants, or use in custom workflows. They work with ChatGPT, GitHub Copilot, browser-based AI tools, and custom LLM apps.

Examples

svar-svelte-grid.md

context for the Svelte grid package

svar-svelte-all.md

context for all packages in one file

Choose the package that best matches your use case. Smaller files provide more focused context and lower token usage, while the "all components" packages provide complete library coverage.

React logo

React

all components 279kb; ~77k tokens
comments 20kb; ~5.6k tokens
core 107kb; ~30k tokens
editor 28kb; ~7.6k tokens
filemanager 33kb; ~9k tokens
filter 49kb; ~13k tokens
gantt 38kb; ~10k tokens
grid 37kb; ~10k tokens
layout 19kb; ~5.3k tokens
menu 21kb; ~5.9k tokens
tasklist 16kb; ~4.4k tokens
toolbar 21kb; ~5.8k tokens
Svelte logo

Svelte

all components 263kb; ~73k tokens
comments 19kb; ~5.4k tokens
core 101kb; ~28k tokens
editor 26kb; ~7.3k tokens
filemanager 32kb; ~8.7k tokens
filter 47kb; ~13k tokens
gantt 36kb; ~9.9k tokens
grid 35kb; ~9.8k tokens
layout 18kb; ~5k tokens
menu 20kb; ~5.6k tokens
tasklist 15kb; ~4.2k tokens
toolbar 20kb; ~5.6k tokens
Vue logo

Vue

all components 275kb; ~76k tokens
comments 20kb; ~5.6k tokens
core 105kb; ~29k tokens
editor 27kb; ~7.5k tokens
filemanager 32kb; ~8.9k tokens
filter 49kb; ~13k tokens
gantt 37kb; ~10k tokens
grid 37kb; ~10k tokens
layout 19kb; ~5.2k tokens
menu 21kb; ~5.8k tokens
tasklist 16kb; ~4.3k tokens
toolbar 21kb; ~5.8k tokens

Choosing the Right Integration

SVAR supports multiple ways of integrating with AI coding assistants. Depending on the capabilities of your development environment, you can use native integrations (Skills and MCP) or fallback to portable context packs.

AI Tool Skills MCP Context Packs
Cursor
Claude Code
Codex CLI
Antigravity CLI
GitHub Copilot Partial
ChatGPT Partial
One-off prompts