Agent-ready Apple developer docs
For Claude Code, Cursor, Copilot, and any MCP client.
Connect your client
Run the following command in your terminal:
claude mcp add --transport http apple-docs https://xdocs.dev/mcpAdd an API key Optional, removes rate limiting
claude mcp add --transport http apple-docs https://xdocs.dev/mcp \
--header "Authorization: Bearer YOUR_API_KEY"⭐️ Add a SKILL.md file Recommended, teaches Claude when to use apple-docs
mkdir -p .claude/skills/searching-apple-docs
Then create .claude/skills/searching-apple-docs/SKILL.md with the content below, or download SKILL.md:
Loading SKILL.md…For more information, see the Claude Code MCP documentation.
Add a custom connector in Claude Desktop settings:
1. Open Claude Desktop 2. Navigate to Settings → Connectors 3. Select "Add custom connector" 4. Enter apple-docs as the name 5. Set URL to https://xdocs.dev/mcp 6. Click "Add"
⭐️ Add a custom skill Recommended, teaches Claude when to use apple-docs
Create a skill folder and package it as a ZIP:
mkdir -p apple-docs
Create apple-docs/SKILL.md with the content below (or download SKILL.md), then ZIP the folder:
Loading SKILL.md…zip -r apple-docs.zip apple-docs/Upload via Settings → Capabilities → Custom Skills. See the Claude Skills documentation for more.
Add to your Cursor MCP settings:
{
"apple-docs": {
"url": "https://xdocs.dev/mcp"
}
}
Create .vscode/mcp.json in your workspace:
{
"servers": {
"apple-docs": {
"type": "http",
"url": "https://xdocs.dev/mcp"
}
}
}Add to your Copilot for Xcode extensions configuration:
{
"mcpServers": {
"apple-docs": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://xdocs.dev/mcp"]
}
}
}
For clients supporting Streamable HTTP, connect directly to the server URL. For stdio-only clients, use mcp-remote as a proxy:
{
"mcpServers": {
"apple-docs": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://xdocs.dev/mcp"]
}
}
}Full control over search and retrieval
Token-aware responses, framework filtering, and complete documentation retrieval.
framework, platform, token_budget, summary_mode, and relevance_threshold options.Pre-indexed and always instant
Unlike on-demand fetching, we've already done the work. Updated twice a week, every search is instant.
Meilisearch Powered
Instant results from a pre-built search index. No fetching required.
Agent Optimized
Token budgets, relevance scoring, and summary modes built for AI context windows.
Complete Coverage
390+ frameworks, Human Interface Guidelines, Swift Language Guide, and MLX docs.
No Snooping
Zero query logging. No middleware. Your searches stay between you and your agent.