AI Integration
SmoothUI is built for AI-assisted development. Discover components via MCP, REST API, and machine-readable catalogs.
Last updated: March 7, 2026
AI Integration
SmoothUI is designed from the ground up for AI-assisted development. Whether you are an AI agent selecting components programmatically, or a developer using AI coding tools, SmoothUI provides first-class integration paths.
Quick Start
Choose the integration that fits your workflow:
| You are... | What to use | Setup |
|---|---|---|
| AI coding agent (Claude, Cursor, Copilot) | MCP Server | npx shadcn@latest mcp init |
| Building custom tooling | REST API | Call https://smoothui.dev/api/v1/... |
| LLM / RAG pipeline | Machine-readable catalog | Fetch https://smoothui.dev/llms-full.txt |
For AI Agents
AI agents can discover, search, and install SmoothUI components through multiple channels.
MCP Server (Recommended)
The fastest way to give an AI assistant full access to SmoothUI. The shadcn MCP server works out of the box with SmoothUI's registry.
npx shadcn@latest mcp init --client claudenpx shadcn@latest mcp init --client cursornpx shadcn@latest mcp init --client vscodeOnce configured, your AI assistant can discover, search, and install any SmoothUI component. See the full MCP Server guide for detailed setup and example prompts.
REST API
For agents that need structured JSON data, SmoothUI provides a public REST API with no authentication required.
curl "https://smoothui.dev/api/v1/suggest?need=animated+tab+navigation"curl "https://smoothui.dev/api/v1/components/animated-tabs?include=source"Key endpoints:
| Endpoint | Purpose |
|---|---|
GET /api/v1/components | List all components with filtering |
GET /api/v1/components/{name} | Get component details and source |
GET /api/v1/components/search?q=... | Keyword search |
GET /api/v1/suggest?need=... | Natural-language component suggestions |
GET /api/v1/blocks | List pre-built page sections |
GET /openapi.json | Full OpenAPI 3.1 specification |
See the full REST API reference for query parameters, response schemas, and error handling.
Machine-Readable Catalog
For LLM context windows and RAG pipelines, SmoothUI provides machine-readable component catalogs following the llms.txt convention:
| URL | Description |
|---|---|
/llms.txt | Compact overview of all components |
/llms-full.txt | Full catalog with metadata, props, and usage hints |
/llms-components.json | Structured JSON catalog for programmatic use |
Agent Workflow
Here is the recommended workflow for AI agents integrating SmoothUI components:
- Discover — Use
/api/v1/suggest?need=...or the MCP server to find relevant components based on what the user needs. - Inspect — Fetch full metadata and source with
/api/v1/components/{name}?include=sourceto understand props and usage. - Install — Use the
installCommandfrom the metadata:npx shadcn@latest add @smoothui/{name}. - Integrate — Use
compositionHintsand the source code to wire the component into the project correctly.