Shadcn MCP for JetBrains AI Assistant

One MCP URL gives JetBrains AI Assistant live tools to search, preview, and install every block, component, and icon on shadcn.io — no hallucinated props, no broken installs.

Install

In your JetBrains IDE (IntelliJ, WebStorm, PyCharm, etc.), open Settings → Tools → AI Assistant → Model Context Protocol (MCP), click + Add, select the HTTP tab, and paste this configuration:

Click Apply once signed in with an active Pro subscription — the config carries your personal token, so treat the full URL like an API key. Pro starts at $19/mo and includes MCP access across every editor plus 6,000+ production-ready blocks. Not Pro yet? See what you unlock →

JetBrains AI Assistant shares MCP configuration across every JetBrains product signed into the same account — once shadcnio is added in IntelliJ, WebStorm and the rest pick it up automatically if Settings Sync is enabled.

Share with your team

JetBrains stores MCP configuration in IDE settings rather than a committed file, so there's no .json-in-repo sharing pattern. Two options:

1. Settings Sync: enable via File → Manage IDE Settings → Settings Sync with the Tools category included. Teammates on the same JetBrains account sync get the MCP entry for free.
2. Per-teammate setup: share the URL pattern (https://www.shadcn.io/api/mcp?token=YOUR_TOKEN) and let each member paste in their own token from /dashboard/account.

A Team or Organization plan gives every workspace member their own Pro entitlement.

Install your first block

use shadcnio to install hero-announcement into my project

What AI Assistant actually does, in order:

1. search_items({ query: "hero-announcement" }) to confirm the slug resolves.
2. get_item({ name: "hero-announcement" }) for a cheap metadata peek.
3. get_install_command({ name: "hero-announcement" }) — returns a shadcn add https://www.shadcn.io/r/hero-announcement.json?token=… URL with your token already in the query string.
4. Runs the install through the IDE's built-in terminal. Files land under components/ui/.

Name the server explicitly ("use shadcnio…") when multiple MCPs are active. Without it, AI Assistant picks one at random or skips tools entirely.

The example above uses hero-announcement, a Pro block — the install runs once you're signed in with an active subscription.

Manage the connection

Return to Settings → Tools → AI Assistant → Model Context Protocol (MCP) to see every configured server. Each row has Edit, Delete, and an enable/disable checkbox. Changes take effect on Apply — no IDE restart needed.

If your Pro subscription lapses, the next tool call returns 403 and AI Assistant marks shadcnio as failed in the MCP panel. Resubscribing restores access on the following request — no reconfigure, no reinstall. The token itself stays valid across subscription gaps.

FAQ